index.json

[{"authors":["aakash"],"categories":null,"content":"Hi! I’m an ECE undergrad at IIT Roorkee and keenly interested in programming , Deep Learning, Machine Learning . My primary areas of interest are Computer Vision, Natural Language Processing . Hit me up any time for a conversation about Lawn Tennis or DL/ML!!!\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"0ef776759294ded106dad9e7bac41096","permalink":"https://vlgiitr.github.io/author/aakash-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/aakash-gupta/","section":"authors","summary":"Hi! I’m an ECE undergrad at IIT Roorkee and keenly interested in programming , Deep Learning, Machine Learning . My primary areas of interest are Computer Vision, Natural Language Processing .","tags":null,"title":"Aakash Gupta","type":"authors"},{"authors":["aakash"],"categories":null,"content":"I am your neighbourhood friendly Pogo\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"1004809bc97ae9da13b82f39a01a2b24","permalink":"https://vlgiitr.github.io/author/abhinav-kumar/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/abhinav-kumar/","section":"authors","summary":"I am your neighbourhood friendly Pogo","tags":null,"title":"Abhinav Kumar","type":"authors"},{"authors":["ayush"],"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"bd1b4d9f6c7d2c965bfad398df732216","permalink":"https://vlgiitr.github.io/author/ayush-singh/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/ayush-singh/","section":"authors","summary":"","tags":null,"title":"Ayush Singh","type":"authors"},{"authors":["aakash"],"categories":null,"content":"Hello, I\u0026rsquo;m Rajdeep, a sophomore from Mathematics and Computing at IIT Roorkee, with a deep passion for Deep Learning and Pure Mathematics.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"908a0db9527619b97c150f4159489544","permalink":"https://vlgiitr.github.io/author/rajdeep-aher/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/rajdeep-aher/","section":"authors","summary":"Hello, I\u0026rsquo;m Rajdeep, a sophomore from Mathematics and Computing at IIT Roorkee, with a deep passion for Deep Learning and Pure Mathematics.","tags":null,"title":"Rajdeep Aher","type":"authors"},{"authors":["siddarth"],"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"2ff8f44667e1f460a28ef05f7414f2d3","permalink":"https://vlgiitr.github.io/author/siddarth-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/siddarth-gupta/","section":"authors","summary":"","tags":null,"title":"Siddarth Gupta","type":"authors"},{"authors":["vansh"],"categories":null,"content":"Hey! Currently pursuing my Bachelor of Technology (BTech) in Engineering Physics at IIT Roorkee. Focused on AI and LLM alignment research along with some work in exploring the multimodal LLMs. Interested in various fields in physics as well and have been working on some projects that explore the intersection of these fields.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"bf71d9e3e426e33cf014f5760946fb60","permalink":"https://vlgiitr.github.io/author/vansh-agrawal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vansh-agrawal/","section":"authors","summary":"Hey! Currently pursuing my Bachelor of Technology (BTech) in Engineering Physics at IIT Roorkee. Focused on AI and LLM alignment research along with some work in exploring the multimodal LLMs. Interested in various fields in physics as well and have been working on some projects that explore the intersection of these fields.","tags":null,"title":"Vansh Agrawal","type":"authors"},{"authors":["vidit"],"categories":null,"content":"Hi! Currently an undergrad in Data Science and AI (DSAI) branch at IIT Roorkee. An avid reader with a deep passion for learning. Currently exploring various fields in AI/ML such as AI security, ML Compilation etc.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"efc0ad39fe9049e9bb7c87aefefdd786","permalink":"https://vlgiitr.github.io/author/vidit-aggarwal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vidit-aggarwal/","section":"authors","summary":"Hi! Currently an undergrad in Data Science and AI (DSAI) branch at IIT Roorkee. An avid reader with a deep passion for learning. Currently exploring various fields in AI/ML such as AI security, ML Compilation etc.","tags":null,"title":"Vidit Aggarwal","type":"authors"},{"authors":["aarush"],"categories":null,"content":"Biography I am a CS graduate from IIT Roorkee and am currently working at Rephrase.ai developing neural networks to generate realistic videos from text.\nVisit my WebPage: https://aarushgupta.github.io\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"59c98840df99e7adee506d2cb226f71e","permalink":"https://vlgiitr.github.io/author/aarush-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/aarush-gupta/","section":"authors","summary":"Biography I am a CS graduate from IIT Roorkee and am currently working at Rephrase.ai developing neural networks to generate realistic videos from text.\nVisit my WebPage: https://aarushgupta.github.io","tags":null,"title":"Aarush Gupta","type":"authors"},{"authors":["aayan"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"4165aae09c5389bc45e00c5a87b47ae9","permalink":"https://vlgiitr.github.io/author/aayan-yadav/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/aayan-yadav/","section":"authors","summary":" ","tags":null,"title":"Aayan Yadav","type":"authors"},{"authors":["barvin"],"categories":null,"content":"Biography Computer Science masters IIIT Hyderabad; bachelors IIT Roorkee; ML apprentice\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"bb65aa82b0cf85d50de98c9aaec1c423","permalink":"https://vlgiitr.github.io/author/abhinaba-bala/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/abhinaba-bala/","section":"authors","summary":"Biography Computer Science masters IIIT Hyderabad; bachelors IIT Roorkee; ML apprentice","tags":null,"title":"Abhinaba Bala","type":"authors"},{"authors":["abhishek"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"8a3ec315d9725239deabb0d05baec1a2","permalink":"https://vlgiitr.github.io/author/abhishek-sinha/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/abhishek-sinha/","section":"authors","summary":" ","tags":null,"title":"Abhishek Sinha","type":"authors"},{"authors":["abya"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"6b5d7ecde6791399968ef979d2f2c54e","permalink":"https://vlgiitr.github.io/author/abya-singh/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/abya-singh/","section":"authors","summary":" ","tags":null,"title":"Abya Singh","type":"authors"},{"authors":["ap"],"categories":null,"content":"Biography My area of research is at the intersection of machine learning, computer vision and robotics. In particular, I am interested in developing perception driven robotic systems for robust navigation in real environments. Previously, I have worked with amazing researchers at Naver Clova AI Research, Adobe Research, Video Analytics Lab, Indian Institute of Science and Cornell Tech.\nVisit my WebPage: https://ap229997.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"19bc883b12cfb922873ad0fd9e342114","permalink":"https://vlgiitr.github.io/author/aditya-prakash/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/aditya-prakash/","section":"authors","summary":"Biography My area of research is at the intersection of machine learning, computer vision and robotics. In particular, I am interested in developing perception driven robotic systems for robust navigation in real environments.","tags":null,"title":"Aditya Prakash","type":"authors"},{"authors":["aniket"],"categories":null,"content":"Biography Music Lover with a weird sense of humour (cringey enough to laugh in pity). Highly interested in mathematical and theoretical aspects of Deep Learning, and also to solve general CV problems in the field of Graphical Deep Learning, Generative Modelling, etc. Would mostly be seen watching The Office in my free time, and arguing why it is way better than Friends :)\nInterests Deep Learning Graphical Deep Learning Computer Vision Generative Modelling ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"d7dac0ed6020a58f8e8e50b2cc65e6d9","permalink":"https://vlgiitr.github.io/author/aniket-agarwal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/aniket-agarwal/","section":"authors","summary":"Biography Music Lover with a weird sense of humour (cringey enough to laugh in pity). Highly interested in mathematical and theoretical aspects of Deep Learning, and also to solve general CV problems in the field of Graphical Deep Learning, Generative Modelling, etc.","tags":null,"title":"Aniket Agarwal","type":"authors"},{"authors":["arpit"],"categories":null,"content":"Biography I am a mechanical engineering student, but I really find my mechanics class the best place to sleep. I like solving real-world problems based on machine and deep learning, but research is not where I belong. Apart from this, I could be seen streaming Netflix or scrolling Instagram.\nInterests Deep Learning Machine Learning ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"51f3a18dd1f6867c37de9a4c5e118a30","permalink":"https://vlgiitr.github.io/author/arpit-khandelwal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/arpit-khandelwal/","section":"authors","summary":"Biography I am a mechanical engineering student, but I really find my mechanics class the best place to sleep. I like solving real-world problems based on machine and deep learning, but research is not where I belong.","tags":null,"title":"Arpit Khandelwal","type":"authors"},{"authors":["chaubey"],"categories":null,"content":"Biography I am a tiny little human being interested in playing with deep neural networks (and a lot of other things). Specific areas of interest include active learning, weak learning and adversarial perturbations. In my free time, I like listening to music and watching tech videos on youtube. I have been a part of VLG since September 2018 and the journey has been wonderful since then. \u0026lt;3\nVisit my WebPage: www.ashutoshchaubey.in\nInterests Deep Learning Active Learning Adversarial ML Weak Learning Web Development ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"268c3a6667a8a1c7911408a55d277db1","permalink":"https://vlgiitr.github.io/author/ashutosh-chaubey/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/ashutosh-chaubey/","section":"authors","summary":"Biography I am a tiny little human being interested in playing with deep neural networks (and a lot of other things). Specific areas of interest include active learning, weak learning and adversarial perturbations.","tags":null,"title":"Ashutosh Chaubey","type":"authors"},{"authors":["mangal"],"categories":null,"content":"Otaku, interested in learning new stuff about anything and everything. Loves anime and good music more than anything. Current interests involve Computer Vision, Robotics, Finance and business management. Wants to open something of his own somewhere along the road.\nVisit my webpage : https://ayushtues.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"64512bf9a40e0b88a349ff1acd7021a0","permalink":"https://vlgiitr.github.io/author/ayush-mangal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/ayush-mangal/","section":"authors","summary":"Otaku, interested in learning new stuff about anything and everything. Loves anime and good music more than anything. Current interests involve Computer Vision, Robotics, Finance and business management. Wants to open something of his own somewhere along the road.","tags":null,"title":"Ayush Mangal","type":"authors"},{"authors":["ayushman"],"categories":null,"content":"Biography An inquisitive computer science enthusiast with a knack for algorithms specially pertaining to Deep Learning and Development. Always in for a hangout over a steaming cup of coffee on a pleasent rainy eve (Potterheads are more welcome).\nVisit my WebPage: https://www.ayushmantripathy.com/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"0deca762182855ef8aec0402b32deb23","permalink":"https://vlgiitr.github.io/author/ayushman-tripathy/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/ayushman-tripathy/","section":"authors","summary":"Biography An inquisitive computer science enthusiast with a knack for algorithms specially pertaining to Deep Learning and Development. Always in for a hangout over a steaming cup of coffee on a pleasent rainy eve (Potterheads are more welcome).","tags":null,"title":"Ayushman Tripathy","type":"authors"},{"authors":["dakshit"],"categories":null,"content":"Biography Hi! I am Dakshit, currently a Research Engineer at Rephrase.ai where I get to work on a cool AI tech which lets you convert text to videos! I recently graduated from IIT Roorkee with a major in Computer Science.\nDuring my undergraduate years, I was fortunate to be the Co-President of VLG, the Treasurer of ACM IIT Roorkee Student Chapter, and a member of the Mobile Development Group. I also got an opportunity to visit some great places to do research in AI, namely, Prof. Katerina Fragkiadaki\u0026rsquo;s group at Carnegie Mellon University, Prof. Venkatesh Babu\u0026rsquo;s group at Indian Institute of Science, and Prof. Marco Pedersoli\u0026rsquo;s group at École de Technologie Supérieure.\nBesides the everyday bustle of my job, I juggle my time between playing the guitar, managing my financial portfolio, reading novels, watching TV series, or just hanging out with my family and friends. Feel free to get in touch if you wanna have a chat!\nVisit my WebPage: https://dakshitagrawal97.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"42b0e125f2f59e141e5c43f1c9804764","permalink":"https://vlgiitr.github.io/author/dakshit-agrawal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/dakshit-agrawal/","section":"authors","summary":"Biography Hi! I am Dakshit, currently a Research Engineer at Rephrase.ai where I get to work on a cool AI tech which lets you convert text to videos! I recently graduated from IIT Roorkee with a major in Computer Science.","tags":null,"title":"Dakshit Agrawal","type":"authors"},{"authors":["dhruv"],"categories":null,"content":"Biography Enter a forest, explore it, find the way out till you find the ripest fruit : Wondering Exploration. I am Chemical Engineering undergrad at IIT Roorkee. Now, keeping exploration aside, I love to play guitar and travel a lot.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"7f6a5d5045bf4a85988fd96f76cf01fc","permalink":"https://vlgiitr.github.io/author/dhruv-grover/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/dhruv-grover/","section":"authors","summary":"Biography Enter a forest, explore it, find the way out till you find the ripest fruit : Wondering Exploration. I am Chemical Engineering undergrad at IIT Roorkee. Now, keeping exploration aside, I love to play guitar and travel a lot.","tags":null,"title":"Dhruv Grover","type":"authors"},{"authors":["divyam"],"categories":null,"content":"Biography I enjoy using mathematics and computational tools in applied settings to solve challenging problems of scientific and societal value. I am passionate about research in embodied agents and multimodal deep learning. I also happen to be a football fanatic and a music geek. Unpopular opinion - post Syd Pink Floyd is miles better than with Syd :)\nVisit my website at : https://dv-fenix.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"a685dff10520da9d565e50178aa180da","permalink":"https://vlgiitr.github.io/author/divyam-goel/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/divyam-goel/","section":"authors","summary":"Biography I enjoy using mathematics and computational tools in applied settings to solve challenging problems of scientific and societal value. I am passionate about research in embodied agents and multimodal deep learning.","tags":null,"title":"Divyam Goel","type":"authors"},{"authors":["hardik"],"categories":null,"content":"Biography My research focuses on Natural Language Processing. Currently, I am a Research engineer at Exawizards Inc, Tokyo. Previously, I have done research assistantship under the supervision of Prof. Pushpak Bhattacharyya where I worked on projects ranging from Multimodal Dialogue System to Multi-hop Question generation.\nVisit my WebPage: https://hardik2396.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"2c8d564d6a308edad69fcfbcef5e8a80","permalink":"https://vlgiitr.github.io/author/hardik-chauhan/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/hardik-chauhan/","section":"authors","summary":"Biography My research focuses on Natural Language Processing. Currently, I am a Research engineer at Exawizards Inc, Tokyo. Previously, I have done research assistantship under the supervision of Prof. Pushpak Bhattacharyya where I worked on projects ranging from Multimodal Dialogue System to Multi-hop Question generation.","tags":null,"title":"Hardik Chauhan","type":"authors"},{"authors":["harsh"],"categories":null,"content":"Biography I am currently a student at Indian Institute of Technology Roorkee majoring in the field of applied mathematics. I enjoy programming and am interested in Machine Learning, Computer security and Cryptocurrencies. And in my free time I read.\nVisit my WebPage: https://chanbong.github.io\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"3482ac6bc0b7aedc2210b6b9168da5fb","permalink":"https://vlgiitr.github.io/author/harsh-kumar/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/harsh-kumar/","section":"authors","summary":"Biography I am currently a student at Indian Institute of Technology Roorkee majoring in the field of applied mathematics. I enjoy programming and am interested in Machine Learning, Computer security and Cryptocurrencies.","tags":null,"title":"Harsh Kumar","type":"authors"},{"authors":["harshil"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"3954f44859c51186139f7049de387806","permalink":"https://vlgiitr.github.io/author/harshil-sajan/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/harshil-sajan/","section":"authors","summary":" ","tags":null,"title":"Harshil Sajan","type":"authors"},{"authors":["harshit"],"categories":null,"content":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers. The group has since then evolved into a group that focuses on both the theoretical and practical knowledge of Deep Learning. Theoretical aspects are covered in campus-open discussions and brainstorming sessions of recent and renowned Deep Learning papers, while practical applications of ideas include research projects and and paper implementations, so as to have a robust understanding of the field. The group works towards developing the research capabilities of the students in the campus, so that students wanting to pursue research after their degree have a strong foundation to stand on.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"cc3cdb30ad1bc3009e4a66886d557514","permalink":"https://vlgiitr.github.io/author/harshit-sharma/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/harshit-sharma/","section":"authors","summary":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers.","tags":null,"title":"Harshit Sharma","type":"authors"},{"authors":["himanshi"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"6375b397a8245406b473c0411c680630","permalink":"https://vlgiitr.github.io/author/himanshi-tibrewal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/himanshi-tibrewal/","section":"authors","summary":" ","tags":null,"title":"Himanshi Tibrewal","type":"authors"},{"authors":["huazib"],"categories":null,"content":"Biography Hey! I am Huzaib, currently a pre-final yearite at IIT Roorkee. I\u0026rsquo;m interested in Generative modeling, natural language processing, self-supervised learning, and the list is getting longer with time. I do not believe in making long-term plans and hence I\u0026rsquo;m not sure where I\u0026rsquo;ll end up but I do keep the fight clean😋.\nTurned into a serious MMA fan from the last couple of years. I\u0026rsquo;m always open to having UFC discussions. I play cricket casually with friends and love mind-wobbling movies. Also addicted to Michael Jackson\u0026rsquo;s pop somehow😝.\nInterests Deep Learning Natural Language Processing Computer Vision Self-Supervised Learning\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"08f33d359ac7f002035a95dc4162408d","permalink":"https://vlgiitr.github.io/author/huzaib-ul-hassan/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/huzaib-ul-hassan/","section":"authors","summary":"Biography Hey! I am Huzaib, currently a pre-final yearite at IIT Roorkee. I\u0026rsquo;m interested in Generative modeling, natural language processing, self-supervised learning, and the list is getting longer with time. I do not believe in making long-term plans and hence I\u0026rsquo;m not sure where I\u0026rsquo;ll end up but I do keep the fight clean😋.","tags":null,"title":"Huzaib Ul Hassan","type":"authors"},{"authors":["ishan"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"4b31841e2ffd23027acab0b123e391cd","permalink":"https://vlgiitr.github.io/author/ishan-garg/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/ishan-garg/","section":"authors","summary":" ","tags":null,"title":"Ishan Garg","type":"authors"},{"authors":["jitesh"],"categories":null,"content":"Biography I am a diligent Deep Learning enthusiast, diving deep into the world of Artificial Intelligence, trying to find solutions to the unanswered problems in the field of AI to help humanity.Apart from my interest in Deep Learning Research, I am also interested in Flutter App Development.\nTo relax my mind amidst the madness that goes around the world, I listen to pop music and read books (mainly philosphical). I am one of those people who follows cricket as a religion. Do contact me if you want to discuss about pop-music, a little philosphy or anything about cricket.\nVisit my WebPage: https://praeclarumjj3.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"131cb2b40c815104c5e7df78f33db59a","permalink":"https://vlgiitr.github.io/author/jitesh-jain/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/jitesh-jain/","section":"authors","summary":"Biography I am a diligent Deep Learning enthusiast, diving deep into the world of Artificial Intelligence, trying to find solutions to the unanswered problems in the field of AI to help humanity.","tags":null,"title":"Jitesh Jain","type":"authors"},{"authors":["kaaira"],"categories":null,"content":"Biography Exploring different domains of Computer Science and particularly interested in Deep Learning, specifically NLP, and backend development. Reading finance and philosophy on the side as well. Also, a voracious reader with an insatiable hunger for classics.\nVisit my WebPage: https://kaairagupta.github.io/\nInterests Android and cross platform application development Natural Language Processing Backend Development ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"c5401e5a9dfcff1eafc6e01260e0176b","permalink":"https://vlgiitr.github.io/author/kaaira-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/kaaira-gupta/","section":"authors","summary":"Biography Exploring different domains of Computer Science and particularly interested in Deep Learning, specifically NLP, and backend development. Reading finance and philosophy on the side as well. Also, a voracious reader with an insatiable hunger for classics.","tags":null,"title":"Kaaira Gupta","type":"authors"},{"authors":["kd"],"categories":null,"content":"Biography I am a first year CS PhD student at the University of Michigan, advised by Justin Johnson. I am broadly interested in deep learning, with applications in computer vision and natural language processing.\nVisit my WebPage: https://kdexd.github.io/\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"ebce41d6363a572f4ecdb1f5ecf3bbce","permalink":"https://vlgiitr.github.io/author/karan-desai/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/karan-desai/","section":"authors","summary":"Biography I am a first year CS PhD student at the University of Michigan, advised by Justin Johnson. I am broadly interested in deep learning, with applications in computer vision and natural language processing.","tags":null,"title":"Karan Desai","type":"authors"},{"authors":["kavya"],"categories":null,"content":"Biography Deep Learning Enthusiast with interest in Probability and Statistics, Machine learning, and problem solving. Loves to learn new stuffs, and always ready for adventures, and listening stories. Think of me as an agent learning slowly the tricks of this open world!\nInterests Deep Learning Machine learning Probability and Statistics ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"edf9aa2e80dee1fdbd2040f035bb11f4","permalink":"https://vlgiitr.github.io/author/kavya-barnwal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/kavya-barnwal/","section":"authors","summary":"Biography Deep Learning Enthusiast with interest in Probability and Statistics, Machine learning, and problem solving. Loves to learn new stuffs, and always ready for adventures, and listening stories. Think of me as an agent learning slowly the tricks of this open world!","tags":null,"title":"Kavya Barnwal","type":"authors"},{"authors":["keerat"],"categories":null,"content":"Biography Researcher-in-the-making. Interested in data-driven decision making to make the world a better place. Passionate debater and Badminton enthusiast. Busy bee but not a workaholic :)\nVisit my webpage : https://kkguliani.netlify.app/\nInterests Deep Learning Machine Intelligence AI for Healthcare Computer Vision ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"7a0a56b8f53e8eae99001ce754ec1b12","permalink":"https://vlgiitr.github.io/author/keerat-kaur-guliani/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/keerat-kaur-guliani/","section":"authors","summary":"Biography Researcher-in-the-making. Interested in data-driven decision making to make the world a better place. Passionate debater and Badminton enthusiast. Busy bee but not a workaholic :)\nVisit my webpage : https://kkguliani.","tags":null,"title":"Keerat Kaur Guliani","type":"authors"},{"authors":["devesh"],"categories":null,"content":"Biography Hi! I am a CS undergrad at IIT Roorkee. My interests include finance, math, and dl exploring Computer Vision Problems. Inquisitive about how things work. In my free time, I watch random youtube recommendations or listen to music.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"3fa58640d62aa759edc05d012fd3ca19","permalink":"https://vlgiitr.github.io/author/kumar-devesh/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/kumar-devesh/","section":"authors","summary":"Biography Hi! I am a CS undergrad at IIT Roorkee. My interests include finance, math, and dl exploring Computer Vision Problems. Inquisitive about how things work. In my free time, I watch random youtube recommendations or listen to music.","tags":null,"title":"Kumar Devesh","type":"authors"},{"authors":["manav"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"28fa0b02f1c9e0507bee19af987137fa","permalink":"https://vlgiitr.github.io/author/manav-goyal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/manav-goyal/","section":"authors","summary":" ","tags":null,"title":"Manav goyal","type":"authors"},{"authors":["mansi"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"182d4b48ae30a410e12cdfe57b0212ae","permalink":"https://vlgiitr.github.io/author/mansi-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/mansi-gupta/","section":"authors","summary":" ","tags":null,"title":"Mansi Gupta","type":"authors"},{"authors":["nikhar"],"categories":null,"content":"\r","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"87f5bc74563d691d861251c14fe259dc","permalink":"https://vlgiitr.github.io/author/manyana-tiwari/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/manyana-tiwari/","section":"authors","summary":"\r","tags":null,"title":"Manyana Tiwari","type":"authors"},{"authors":["nikhar"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"22447a7439aa3ca42163ccbc47c58cc8","permalink":"https://vlgiitr.github.io/author/nikhar-waghela/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/nikhar-waghela/","section":"authors","summary":" ","tags":null,"title":"Nikhar Waghela","type":"authors"},{"authors":["nikhil"],"categories":null,"content":"Biography I like to explore Artifical Intelligence and Human Behaviours Analysis. Besides that I am trying to maintain a work life balance where I can eat without gaining weight.\nInterests Deep Learning Artificial Intelligence Human Behaviour Analysis ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"4004cb550868341d3a3c5d2c94ce8aae","permalink":"https://vlgiitr.github.io/author/nikhil-agrawal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/nikhil-agrawal/","section":"authors","summary":"Biography I like to explore Artifical Intelligence and Human Behaviours Analysis. Besides that I am trying to maintain a work life balance where I can eat without gaining weight.\nInterests Deep Learning Artificial Intelligence Human Behaviour Analysis ","tags":null,"title":"Nikhil Agrawal","type":"authors"},{"authors":["nishant"],"categories":null,"content":"Biography Hi! I\u0026rsquo;m into Deep Learning, Machine Learning and Data Analytics as well. My primary areas of interest are Computer Vision, Adversarial ML and core Data Science too. Hit me up any time for a conversation about Basketball or DL/ML!!!\nInterests Deep Learning Machine Learning Data Science Computer Vision ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"7a8fa54751fe9187a07a88957f6859e0","permalink":"https://vlgiitr.github.io/author/nishant-bhansali/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/nishant-bhansali/","section":"authors","summary":"Biography Hi! I\u0026rsquo;m into Deep Learning, Machine Learning and Data Analytics as well. My primary areas of interest are Computer Vision, Adversarial ML and core Data Science too. Hit me up any time for a conversation about Basketball or DL/ML!","tags":null,"title":"Nishant Bhansali","type":"authors"},{"authors":["omkar"],"categories":null,"content":"Biography A versatile research-oriented undergrad student at IIT Roorkee. Passionate about cutting-edge AI with applications in Computer Vision and Natural Language Processing having experience in Production \u0026amp; Industrial Engineering\nLove diving into deep conversations on design, psychology, and spirituality\nInterests Deep Learning UI/UX Design Computer Vision Psychology ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"8a19d9141f17143538cba198f81e1560","permalink":"https://vlgiitr.github.io/author/omkar-bhalero/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/omkar-bhalero/","section":"authors","summary":"Biography A versatile research-oriented undergrad student at IIT Roorkee. Passionate about cutting-edge AI with applications in Computer Vision and Natural Language Processing having experience in Production \u0026amp; Industrial Engineering\nLove diving into deep conversations on design, psychology, and spirituality","tags":null,"title":"Omkar Bhalero","type":"authors"},{"authors":["partha"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"da743d50ca7bc8e692f0bbca1c23fa06","permalink":"https://vlgiitr.github.io/author/partha-kaushik/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/partha-kaushik/","section":"authors","summary":" ","tags":null,"title":"Partha Kaushik","type":"authors"},{"authors":["prakash"],"categories":null,"content":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers. The group has since then evolved into a group that focuses on both the theoretical and practical knowledge of Deep Learning. Theoretical aspects are covered in campus-open discussions and brainstorming sessions of recent and renowned Deep Learning papers, while practical applications of ideas include research projects and and paper implementations, so as to have a robust understanding of the field. The group works towards developing the research capabilities of the students in the campus, so that students wanting to pursue research after their degree have a strong foundation to stand on.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"348940815619cb2ae8f971149522e85c","permalink":"https://vlgiitr.github.io/author/prakash-pandey/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/prakash-pandey/","section":"authors","summary":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers.","tags":null,"title":"Prakash Pandey","type":"authors"},{"authors":["pramod"],"categories":null,"content":"Biography Loves to watch Netflix. Covered a lot of series may we can sometimes discuss abt it a lot. Have a great interest in ML, DL, Data Structures, and Algorithms. I feel mesmerizing to watch my neural networks in training and inaction as well with good accuracies (though the hard part is to code them correctly and confirm that the network is being trained :p).\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"579554484fc51714a537649536b869d3","permalink":"https://vlgiitr.github.io/author/pramod-mehta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/pramod-mehta/","section":"authors","summary":"Biography Loves to watch Netflix. Covered a lot of series may we can sometimes discuss abt it a lot. Have a great interest in ML, DL, Data Structures, and Algorithms. I feel mesmerizing to watch my neural networks in training and inaction as well with good accuracies (though the hard part is to code them correctly and confirm that the network is being trained :p).","tags":null,"title":"Pramod Mehta","type":"authors"},{"authors":["rohan"],"categories":null,"content":"Biography Passionate for Deep learning and Mathematics, Interested in Statistical analysis in deep learning\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"a36473218d1c18bbc57ddacf327f2f16","permalink":"https://vlgiitr.github.io/author/rohan-garg/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/rohan-garg/","section":"authors","summary":"Biography Passionate for Deep learning and Mathematics, Interested in Statistical analysis in deep learning","tags":null,"title":"Rohan Garg","type":"authors"},{"authors":["rusia"],"categories":null,"content":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers. The group has since then evolved into a group that focuses on both the theoretical and practical knowledge of Deep Learning. Theoretical aspects are covered in campus-open discussions and brainstorming sessions of recent and renowned Deep Learning papers, while practical applications of ideas include research projects and and paper implementations, so as to have a robust understanding of the field. The group works towards developing the research capabilities of the students in the campus, so that students wanting to pursue research after their degree have a strong foundation to stand on.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"814e731cc7358d856e724b4bf6ded741","permalink":"https://vlgiitr.github.io/author/sagar-rusia/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/sagar-rusia/","section":"authors","summary":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers.","tags":null,"title":"Sagar Rusia","type":"authors"},{"authors":["sarthak"],"categories":null,"content":"Biography Going deep in learning.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"6f79d5895d93e84e8587a2382a71e325","permalink":"https://vlgiitr.github.io/author/sarthak-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/sarthak-gupta/","section":"authors","summary":"Biography Going deep in learning.","tags":null,"title":"Sarthak Gupta","type":"authors"},{"authors":["mishra"],"categories":null,"content":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers. The group has since then evolved into a group that focuses on both the theoretical and practical knowledge of Deep Learning. Theoretical aspects are covered in campus-open discussions and brainstorming sessions of recent and renowned Deep Learning papers, while practical applications of ideas include research projects and and paper implementations, so as to have a robust understanding of the field. The group works towards developing the research capabilities of the students in the campus, so that students wanting to pursue research after their degree have a strong foundation to stand on.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"398502c675f9f5289a6a5c915d64ce7c","permalink":"https://vlgiitr.github.io/author/saurabh-mishra/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/saurabh-mishra/","section":"authors","summary":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers.","tags":null,"title":"Saurabh Mishra","type":"authors"},{"authors":["shashank"],"categories":null,"content":"Biography A curious computer scientist (in making) and an inquisitive computer engineer (in making as well), exploring the world of research and trying to answer the unanswered in the fascinating field of Deep Learning. Familiar with concepts in Deep Learning, linear algebra, probability and statistics and computer science fundamentals. Always open to discussions on system management, new technologies and Harry Potter.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"592d4144643d44fd170307ff8bfdf288","permalink":"https://vlgiitr.github.io/author/shashank-aital/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/shashank-aital/","section":"authors","summary":"Biography A curious computer scientist (in making) and an inquisitive computer engineer (in making as well), exploring the world of research and trying to answer the unanswered in the fascinating field of Deep Learning.","tags":null,"title":"Shashank Aital","type":"authors"},{"authors":["shivank"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"f2a73559b55e00318e9cd8e2a2682cff","permalink":"https://vlgiitr.github.io/author/shivank-garg/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/shivank-garg/","section":"authors","summary":" ","tags":null,"title":"Shivank Garg","type":"authors"},{"authors":["shanks"],"categories":null,"content":"Biography “One Piece Does Exist!” My research interests include supervised learning and generative models, and I have other experience in computer vision applications. I am passionate about impact-oriented research.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"213b7c924a26393b9651f49d0c295a87","permalink":"https://vlgiitr.github.io/author/shivshankar-shukla/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/shivshankar-shukla/","section":"authors","summary":"Biography “One Piece Does Exist!” My research interests include supervised learning and generative models, and I have other experience in computer vision applications. I am passionate about impact-oriented research.","tags":null,"title":"Shivshankar Shukla","type":"authors"},{"authors":["shubham"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"1b3d99624a46e0588d0602354bb489fc","permalink":"https://vlgiitr.github.io/author/shubham-kumar-singh/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/shubham-kumar-singh/","section":"authors","summary":" ","tags":null,"title":"Shubham Kumar Singh","type":"authors"},{"authors":["shweta"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"df55b7db8b51f76c89c30a679d79b5f8","permalink":"https://vlgiitr.github.io/author/shweta-singh/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/shweta-singh/","section":"authors","summary":" ","tags":null,"title":"Shweta Singh","type":"authors"},{"authors":["soham"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"fb9a9ba1673437a1722f387aec81be56","permalink":"https://vlgiitr.github.io/author/soham-chatterjee/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/soham-chatterjee/","section":"authors","summary":" ","tags":null,"title":"Soham Chatterjee","type":"authors"},{"authors":["tanishq"],"categories":null,"content":"\r","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"67102d50a5ace37f81c3c4fe9ca229a5","permalink":"https://vlgiitr.github.io/author/tanishq-agarwal/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/tanishq-agarwal/","section":"authors","summary":"\r","tags":null,"title":"Tanishq Agarwal","type":"authors"},{"authors":["tanya"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"33e465e5cd7c00ccd62f7e0bbc0075f9","permalink":"https://vlgiitr.github.io/author/tanya-bharti/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/tanya-bharti/","section":"authors","summary":" ","tags":null,"title":"Tanya Bharti","type":"authors"},{"authors":["tushdon"],"categories":null,"content":"Biography A CV and RL Enthusiast\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"81ae16412e64e121a42a9c5f2650cbd7","permalink":"https://vlgiitr.github.io/author/tushar-sahu/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/tushar-sahu/","section":"authors","summary":"Biography A CV and RL Enthusiast","tags":null,"title":"Tushar Sahu","type":"authors"},{"authors":["gosain"],"categories":null,"content":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers. The group has since then evolved into a group that focuses on both the theoretical and practical knowledge of Deep Learning. Theoretical aspects are covered in campus-open discussions and brainstorming sessions of recent and renowned Deep Learning papers, while practical applications of ideas include research projects and and paper implementations, so as to have a robust understanding of the field. The group works towards developing the research capabilities of the students in the campus, so that students wanting to pursue research after their degree have a strong foundation to stand on.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"9575b944b0afb27c6173e47617ff7548","permalink":"https://vlgiitr.github.io/author/vaibhav-gosain/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vaibhav-gosain/","section":"authors","summary":"The Vision and Language Group, ACM IIT Roorkee Chapter, is a student run group which aims to foster an on-campus research-centric Deep Learning community. The group was formed in 2017 to provide a platform to meet and discuss Deep Learning research papers.","tags":null,"title":"Vaibhav Gosain","type":"authors"},{"authors":["preethi"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"f0e6d6817c434d6a721679b2abeb31af","permalink":"https://vlgiitr.github.io/author/vijayapreethi-s-r/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vijayapreethi-s-r/","section":"authors","summary":" ","tags":null,"title":"Vijayapreethi S R","type":"authors"},{"authors":["zenith"],"categories":null,"content":" ","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"37a89df4019f68efec62c9e9b3f172d9","permalink":"https://vlgiitr.github.io/author/zenith-gupta/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/zenith-gupta/","section":"authors","summary":" ","tags":null,"title":"Zenith Gupta","type":"authors"},{"authors":["vipul"],"categories":null,"content":"Biography Diving deep to learn how things are happening at the base level is what I do. Also, I love to celebrate my birthday because before that, I was in 9 months quarantine, and I hate quarantine.\n","date":1589587200,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":1589587200,"objectID":"8c3af16db250d2e3aed705d25e40b162","permalink":"https://vlgiitr.github.io/author/vipul-kumar/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vipul-kumar/","section":"authors","summary":"Biography Diving deep to learn how things are happening at the base level is what I do. Also, I love to celebrate my birthday because before that, I was in 9 months quarantine, and I hate quarantine.","tags":null,"title":"Vipul Kumar","type":"authors"},{"authors":["admin"],"categories":null,"content":"The Vision and Language Group, part of ACM IIT Roorkee Chapter, is a student run group that aims to foster a research-centric Deep Learning Community at IIT Roorkee. We regularly hold open discussions on various DL, CV, NLP papers presented in the latest conferences/journals and also on various general topics pertaining to the Deep Learning field. These discussions are open for anyone to join in.\nApart from this, the group members are also involved in various research based projects, sometimes in collaboration with other professors, with the ultimate goal to bring forth a positive impact in a sub-field we are interested in and also aim for some of the tier-1 conferences.\nWe are constantly looking for new collaborations, so do contact us if you find our work interesting. Also you can follow us up on Facebook and Twitter to receive updates about our activities.\n","date":-62135596800,"expirydate":-62135596800,"kind":"term","lang":"en","lastmod":-62135596800,"objectID":"2525497d367e79493fd32b198b28f040","permalink":"https://vlgiitr.github.io/author/vision-and-language-group/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/author/vision-and-language-group/","section":"authors","summary":"The Vision and Language Group, part of ACM IIT Roorkee Chapter, is a student run group that aims to foster a research-centric Deep Learning Community at IIT Roorkee. We regularly hold open discussions on various DL, CV, NLP papers presented in the latest conferences/journals and also on various general topics pertaining to the Deep Learning field.","tags":null,"title":"Vision and Language Group","type":"authors"},{"authors":null,"categories":[],"content":"Machine unlearning is an emergent subfield of machine learning that aims to remove the influence of a specific subset of training examples — the \u0026ldquo;forget set\u0026rdquo; — from a trained model. Furthermore, an ideal unlearning algorithm would remove the influence of certain examples while maintaining other beneficial properties, such as the accuracy on the rest of the train set and generalization to held-out examples.\nA straightforward way to produce this unlearned model is to retrain the model on an adjusted training set that excludes the samples from the forget set.\n","date":1700956800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1700956800,"objectID":"376258fe95be58232210b8eab05a9871","permalink":"https://vlgiitr.github.io/project/machine_unlearning/","publishdate":"2023-11-26T00:00:00Z","relpermalink":"/project/machine_unlearning/","section":"project","summary":"We aimed to research on removing specific classes of data from a pre-trained LLM model by using Adapter Based approaches and model pruning ","tags":[],"title":"Machine Unlearning","type":"project"},{"authors":null,"categories":[],"content":"While many state-of-the-art LLMs have shown poor logical and basic mathematical reasoning, recent works try to improve their problem-solving abilities using prompting techniques. We propose giving \u0026ldquo;hints\u0026rdquo; to improve the language model’s performance on advanced mathematical problems, taking inspiration from how humans approach math pedagogically. We also test the model’s adversarial robustness to wrong hints. We demonstrate the effectiveness of our approach by evaluating various LLMs, presenting them with a diverse set of problems of different difficulties and topics from the MATH dataset and comparing against techniques such as one-shot, few-shot, and chain of thought prompting.\n","date":1700956800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1700956800,"objectID":"be201a40958b828b7e6a862a28d68a32","permalink":"https://vlgiitr.github.io/project/llm-math/","publishdate":"2023-11-26T00:00:00Z","relpermalink":"/project/llm-math/","section":"project","summary":"A study on enhancing LLM performance in solving math problems through hints, while examining the impact of adversarial prompts.","tags":[],"title":"Give me a hint: Can LLMs take a hint to solve math problems?","type":"project"},{"authors":null,"categories":[],"content":"This study addresses the challenge of machine unlearning in light of growing privacy regulations and the need for adaptable AI systems. We present a novel approach, PruneLoRA where we leverage LoRA to selectively modify a subset of the pruned model’s parameters, thereby reducing the computational cost, memory requirements and improving the model’s ability to retain performance on the remaining classes\n","date":1700956800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1700956800,"objectID":"eff7a2872112504493e31a3f4014962d","permalink":"https://vlgiitr.github.io/project/lora_unlearn/","publishdate":"2023-11-26T00:00:00Z","relpermalink":"/project/lora_unlearn/","section":"project","summary":"LoRA-Unlearn introduces a new Machine Unlearning paradigm, using LoRA to fine-tune sparse models for class unlearning.","tags":[],"title":"LoRA-Unlearn","type":"project"},{"authors":null,"categories":[],"content":"Our study, StegaVision, aims to enhance image steganography by integrating attention mechanisms into an autoencoder-based model. Our approach focuses on dynamically adjusting the importance of different parts of the image through attention mechanisms. This helps in better embedding the hidden information while maintaining the image\u0026rsquo;s visual quality. We specifically explore two types of attention mechanisms—Channel Attention and Spatial Attention—and test their effectiveness on an autoencoder model.\n","date":1700956800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1700956800,"objectID":"5e67682edfd4aa92911e8b6d9386303a","permalink":"https://vlgiitr.github.io/project/stegavision/","publishdate":"2023-11-26T00:00:00Z","relpermalink":"/project/stegavision/","section":"project","summary":"Analysis of importance of different attention mechanism in Image steganography within an auto encoder framework.","tags":[],"title":"StegaVision: Enhancing Steganography with Attention Mechanism","type":"project"},{"authors":null,"categories":[],"content":"An experiment in testing a novel method to train neural networks inspired by the Forward-Forward Algorithm proposed by Geoffrey Hinton by updating weights of a layer by calculating the loss at each intermediate layer instead of backpropagating the losses through the entire network.\nIn the original paper, instead of relying on the traditional forward and backward passes of backpropagation, the method utilized two forward passes — one with positive, real data and the other with negative data. With our modified method we were able to achieve an error rate of less than 2% for a fully connected network and convolutional network on the MNIST dataset.\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"465cf7c24246eb848847db0a37a7c6f9","permalink":"https://vlgiitr.github.io/project/layer-level-loss-optimisation/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/project/layer-level-loss-optimisation/","section":"project","summary":"Experiment to test a method to train neural networks inspired by the Forward-Forward Algorithm ","tags":[],"title":"Layer Level Loss Optimisation - 2023","type":"project"},{"authors":null,"categories":[],"content":"The NeurIPS 2022 The SENSORIUM competition aimed to find the best neural predictive model that can predict the activity of thousands of neurons in the primary visual cortex of mice in response to natural images.\nIn our submission for this competition, we attempted to improve the baseline model for the competition track- Sensorium+, where neural activity was to be predicted with given visual stimuli and other behavioural variables.\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"e35b7447b31dec73623e90900dec827a","permalink":"https://vlgiitr.github.io/project/sensorium/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/project/sensorium/","section":"project","summary":"In the NeurIPS 2022 SENSORIUM competition, we aimed to enhance the baseline model in the Sensorium+ track for predicting mouse primary visual cortex neuron activity based on natural images and behavioral data.","tags":[],"title":"Sensorium 2022","type":"project"},{"authors":null,"categories":[],"content":"The PyTorch codebase for DEAP Cache: Deep Eviction Admission and Prefetching for Cache.\nIn this paper, we propose a DL based approach to tackle the problem of Cache Replacement. This is the first time an approach has tried learning all the three policies: Admission, Prefetching and Eviction. Unlike, previous methods which relied on past statistics for carrying out cache replacement, we predict future statistics (frequency and recency) and then use an online RL-algorithm for eviction.\n","date":1600473600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1600473600,"objectID":"22f651d22fc3c1536e39bc4548dbbe42","permalink":"https://vlgiitr.github.io/project/deap/","publishdate":"2020-09-19T00:00:00Z","relpermalink":"/project/deap/","section":"project","summary":"The PyTorch codebase for DEAP Cache: Deep Eviction Admission and Prefetching for Cache.","tags":[],"title":"Deep Cache Replacement - 2020","type":"project"},{"authors":null,"categories":[],"content":"This repo contains a list of topics which we feel that one should be comfortable with before appearing for a DL interview. This list is by no means exhaustive (as the field is very wide and ever growing).\n","date":1599955200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1599955200,"objectID":"2472f47226c952f53a63abf1d9538735","permalink":"https://vlgiitr.github.io/project/dl_topics/","publishdate":"2020-09-13T00:00:00Z","relpermalink":"/project/dl_topics/","section":"project","summary":"Resources for DL","tags":[],"title":"DL Topics","type":"project"},{"authors":null,"categories":null,"content":"GenZoo is a repository that provides implementations of generative models in various frameworks, namely Tensorflow and Pytorch. This was a project taken up by VLG-IITR for the summers of 2019, done with the collaborative efforts of various students.\n","date":1571529600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1571529600,"objectID":"f0d82672743b7f179a62328d9613f70f","permalink":"https://vlgiitr.github.io/project/genzoo/","publishdate":"2019-10-20T00:00:00Z","relpermalink":"/project/genzoo/","section":"project","summary":"GenZoo is a repository that provides implementations of generative models in various frameworks","tags":null,"title":"GenZoo - 2019","type":"project"},{"authors":null,"categories":[],"content":"This repository contains the code of our model submitted for the ICMI 2018 EmotiW Group-Level Emotion Recognition Challenge. The model was ranked 4th in the challenge. The paper proposes an end-to-end model for jointly learning the scene and facial features of an image for group-level emotion recognition.\n","date":1541462400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1541462400,"objectID":"c87c8b943a3b78bc4178e7513e713c84","permalink":"https://vlgiitr.github.io/project/emoticon/","publishdate":"2018-11-06T00:00:00Z","relpermalink":"/project/emoticon/","section":"project","summary":"Paper Implementation of a end-to-end model for jointly learning the scene and facial features of an image for group-level emotion recognition.","tags":[],"title":"Group-Level-Emotion-Recognition - 2018","type":"project"},{"authors":null,"categories":[],"content":"This repository is a stable Pytorch implementation of a Neural Turing Machine and contains the code for training, evaluating and visualizing results for the Copy, Repeat Copy, Associative Recall and Priority Sort tasks. The code has been tested for all 4 tasks and the results obtained are in accordance with the results mentioned in the paper.\n","date":1537920000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1537920000,"objectID":"6c599be0effd126e8a39b816c3110c0b","permalink":"https://vlgiitr.github.io/project/ntm/","publishdate":"2018-09-26T00:00:00Z","relpermalink":"/project/ntm/","section":"project","summary":"This PyTorch repository provides a reliable implementation of a Neural Turing Machine (NTM) for training, evaluating, and visualizing results across Copy, Repeat Copy, Associative Recall, and Priority Sort tasks, with results matching those reported in the paper.","tags":[],"title":"Neural Turing Machines - 2018","type":"project"},{"authors":null,"categories":null,"content":"This is the Pytorch implementation of the paper Dynamic Memory Network for Visual and Textual Question Answering. This paper is an improved version of the original paper Ask Me Anything: Dynamic Memory Networks for Natural Language Processing. The major difference between these ideas is in the functioning of the input module and the memory module which has been explained in detail in the IPython notebook file of this repo.\n","date":1528416000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1528416000,"objectID":"01c72e5ebbb45d33ff2c28e644f63120","permalink":"https://vlgiitr.github.io/project/dmn_plus/","publishdate":"2018-06-08T00:00:00Z","relpermalink":"/project/dmn_plus/","section":"project","summary":"Pytorch implementation of the paper Dynamic Memory Network for Visual and Textual Question Answering","tags":null,"title":"Dynamic Memory Network Plus - 2018","type":"project"},{"authors":null,"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"f35aefb20d3f3bed558423f9a4841576","permalink":"https://vlgiitr.github.io/blogs/posts/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/blogs/posts/","section":"blogs","summary":"","tags":null,"title":"","type":"blogs"},{"authors":["Shivank Garg","Manyana Tiwari"],"categories":null,"content":"","date":1719532800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1719532800,"objectID":"72d551250ea9c435699ccfd50e31a7f1","permalink":"https://vlgiitr.github.io/publication/tmlr/","publishdate":"2024-06-28T00:00:00Z","relpermalink":"/publication/tmlr/","section":"publication","summary":"In this paper, we extend the study of concept ablation within pre-trained models as introduced in 'Ablating Concepts in Text-to-Image Diffusion Models'. Our work focuses on reproducing the results achieved by the different variants of concept ablation proposed through predefined metrics. We also introduce a novel variant of concept ablation—trademark ablation. This variant combines the principles of memorization and instance ablation to tackle the nuanced influence of proprietary or branded elements in model outputs. Further, our research contributions include an observational analysis of the model's limitations. Moreover, we investigate the model's behavior in response to ablation leakage-inducing prompts, which aim to indirectly ablate concepts, revealing insights into the model's resilience and adaptability. We also observe the model's performance degradation on images generated by concepts far from its target ablation concept.","tags":[],"title":"Unmasking the Veil: An Investigation into Concept Ablation for Privacy and Copyright Protection in Images","type":"publication"},{"authors":["Shweta Singh","Aayan Yadav","Jitesh Jain","Humphrey Shi","Justin Johnson","Karan Desai"],"categories":null,"content":"","date":1711497600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1711497600,"objectID":"a9ebf80e2f67425bb205cb9aa81d460c","permalink":"https://vlgiitr.github.io/publication/eccv/","publishdate":"2024-03-27T00:00:00Z","relpermalink":"/publication/eccv/","section":"publication","summary":"The Common Objects in Context (COCO) dataset has been instrumental in benchmarking object detectors over the past decade. Like every dataset, COCO contains subtle errors and imperfections stemming from its annotation procedure. With the advent of high-performing models, we ask whether these errors of COCO are hindering its utility in reliably benchmarking further progress. In search for an answer, we inspect thousands of masks from COCO (2017 version) and uncover different types of errors such as imprecise mask boundaries, non-exhaustively annotated instances, and mislabeled masks. Due to the prevalence of COCO, we choose to correct these errors to maintain continuity with prior research. We develop COCO-ReM (Refined Masks), a cleaner set of annotations with visibly better mask quality than COCO-2017. We evaluate fifty object detectors and find that models that predict visually sharper masks score higher on COCO-ReM, affirming that they were being incorrectly penalized due to errors in COCO-2017. Moreover, our models trained using COCO-ReM converge faster and score higher than their larger variants trained using COCO-2017, highlighting the importance of data quality in improving object detectors. With these findings, we advocate using COCO-ReM for future object detection research. Our dataset is available at https://cocorem.xyz","tags":[],"title":"Benchmarking Object Detectors with COCO: A New Path Forward","type":"publication"},{"authors":["Abhishek Sinha","Himanshi Tibrewal","Mansi Gupta","Nikhar Waghela","Shivank Garg"],"categories":null,"content":"","date":1708732800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1708732800,"objectID":"530b5a6336c2bedd38a334f0a47c0a02","permalink":"https://vlgiitr.github.io/publication/aaai/","publishdate":"2024-02-24T00:00:00Z","relpermalink":"/publication/aaai/","section":"publication","summary":"In this evolving era of machine learning security, membership inference attacks have emerged as a potent threat to the confidentiality of sensitive data. In this attack, adversaries aim to determine whether a particular point was used during the training of a target model. This paper proposes a new method to gauge a data point's membership in a model's training set. Instead of correlating loss with membership, as is traditionally done, we have leveraged the fact that training examples generally exhibit higher confidence values when classified into their actual class. During training, the model is essentially being 'fit' to the training data and might face particular difficulties in generalization to unseen data. This asymmetry leads to the model achieving higher confidence on the training data as it exploits the specific patterns and noise present in the training data. Our proposed approach leverages the confidence values generated by the machine learning model. These confidence values provide a probabilistic measure of the model's certainty in its predictions and can further be used to infer the membership of a given data point. Additionally, we also introduce another variant of our method that allows us to carry out this attack without knowing the ground truth(true class) of a given data point, thus offering an edge over existing label-dependent attack methods.","tags":[],"title":"Confidence Is All You Need for MI Attacks","type":"publication"},{"authors":[""],"categories":null,"content":"Widely used machine learning algorithms are able to learn from new data using batch or online training methods but are incapable of efficiently adapting to data removal. Why do we need data removal though you might think. Turns out data removal is required to address various issues around privacy, fairness, and data quality. For example, the “Right to be Forgotten” in the European Union’s General Data Protection Regulation (GDPR) provides individuals with the right to request the removal of their data from an organization\u0026rsquo;s records.\nNow comes the main question here. How do we go about deleting, or in better words, unlearning about this data.\nThere are a number of approaches to go about this which are broadly categorized into two segments: exact unlearning and approximate unlearning. Exact unlearning algorithms reduce the large computational cost of naïve retraining by structuring the initial training so as to allow for more efficient retraining; in doing so they replicate the same model that would have been produced under naïve retraining. In contrast, approximate unlearning algorithms avoid the need for full retraining, speeding up the process of unlearning by allowing a degree of approximation between the output model and the naïve retrained model.\nSISA-Sharded Isolated Sliced and Aggregated The SISA algorithm is an exact unlearning algorithm which tries to reduce the time taken in naïve unlearning. This is achieved by a reorganization of the training dataset, known as sharding and slicing.\nThe full SISA algorithm is applicable to any machine learning model that has been trained incrementally, for example, via gradient descent. The loss function for such models need not be strongly convex.\nMethodology\nThe SISA training process to consist of four key steps - Sharded, Isolated, Sliced, and Aggregated. The training data is split into S shards, which are further split into R slices. S independent models are trained incrementally on the slices, and predictions of these models are aggregated to form a final output.\nThe data to unlearn is highlighted in red in this diagram. To unlearn this data point, only M2 needs to be retrained, and this process starts from slice D2,2.\nSHARDING\nThe original training dataset is separated into approximately equal-sized shards, with each training data point contained in exactly one shard.\nISOLATION\nEach of the shards is trained in isolation from the other shards, restricting the influence of each data point to a single shard.\nSLICING\nEach of the shards are sub-divided into slices, which are presented to the algorithm incrementally as training proceeds. The trained model states are saved after each slice.\nAGGREGATION\nTo form the final model prediction for a data point, the predictions of each sharded model are aggregated.\nAlgorithm\nWhenever a removal request for a single data point comes in, only the model trained on the shard containing the particular data point needs to be retrained and, moreover, retraining need only begin from the slice containing the data point. As a result, the expected retraining time is faster compared to naïve retraining; the exact speed-up depends on the number of shards and slices used.\n**Algorithm: Initial training with SISA.** **Input:** training data D, number of shards S, number of slices R, number of epochs for each slice e. **Output:** ensemble of models h = ($h1$, . . . , hS) and intermediary model states h˜ = ({h˜i,0, . . . , h˜i,R})Si=1. 1: **procedure** SisaTrain(D; S, R, e) 2: split the data randomly into shards D1, . . . , DS and save shard indices for each data point 3: split each shard Di randomly into R slices Di,1, . . . Di,R and save slice indices for each data point 4: randomly initialise (h˜1,0, . . . , h˜S,0) 5: **for** i = 1; i ≤ S; i++ **do** 6: **for** j = 1; j ≤ R; j++ **do** 7: hi,j ← Train Di,1 ∪ · · · ∪ Di,j | h˜i,j−1 for ej epochs 8: save model state h˜i,j of model hi,j 9: **end for** 10: hi ← hi,R 11: **end for** 12: **return** h = (h1, . . . , hS), h˜U = ({h˜i,0, . . . , h˜i,R})Si=1. 13: **end procedure** Efficiency\nThe number of shards, S, is an efficiency parameter i.e. increasing the number of shards increases the efficiency of SISA, but will degrade the predictive performance of the resultant machine learning model compared to a lower number of shards. Increasing the number of slices in each shard, R, reduces the retraining time but this does not degrade accuracy, provided that the epochs in training are carefully chosen. However, an increase in R does come at increased storage costs due to the increased number of saved model states. The efficiency-storage trade-off of R may be preferable to the efficiency-effectiveness trade-off of S.\nDaRE Forests This is an unlearning algorithm that is specific to decision-tree and random-forest based machine learning models for binary classification. This is done through the development of Data Removal-Enabled (DaRE) trees, and the ensemble of these to form DaRE Forests (DaRE RF). Through the use of strategic thresholding at decision nodes for continuous attributes, high-level random nodes, and caching certain statistics at all nodes, DaRE trees enable efficient removal of training instances.\nMethodology\nDaRE forest ensembles in the same way as a random forest, in particular a random subset of p features are considered at each split. As in regular decision trees, DaRE trees are trained recursively by selecting, at most nodes, an attribute and threshold that optimizes a split criterion. They differ from regular decision trees in three key ways as follows.\nRandom nodes: The top $d_rmax$ levels of nodes in a DaRE tree are random nodes, where $d_rmax$ is an integer hyperparameter. Threshold sampling: During training and deletion, DaRE trees randomly sample k valid thresholds at any node that is neither a random node nor a leaf. These are thresholds that lie between two adjacent data points with opposite labels. Doing so reduces the amount of statistics one needs to store at each node and speeds up computation. Statistics caching: At each node, for each of the k candidate valid thresholds v, various additional statistics are stored and updated. In each case these statistics are sufficient to recompute the split criterion scores and to determine the validity of the current thresholds. As a result, the removal mechanism is able to recall training data from the stored leaf instances, meaning that training data is not required as an explicit input to the mechanism. **Algorithm: DareTrain(D, 0; drmax, k) trains a single DaRE tree Input**: data Dnode, depth d, random node depth drmax, threshold candidate size k. **Output**: trained subtree rooted at a level-d node. 1: **procedure** DareTrain(Dnode, d; drmax, k) 2: **if** stopping criteria reached **then** 3: node ← LeafNode() 4: save instance counts |Dnode|, |D1| 5: save leaf-instance pointers(node, Dnode) 6: compute leaf value(node) 7: **else** 8: **if** d \u0026lt; drmax **then** 9: node ← RandomNode() 10: save instance counts |Dnode|, |Dnode,1| 11: a ← randomly sample attribute(Dnode) 12: v ← randomly sample threshold ∈ [amin, amax) 13: save threshold statistics(node, Dnode, a, v) 14: **else** 15: node ← GreedyNode() 16: save instance counts |Dnode|, |Dnode,1| 17: A ← randomly sample ˜p attributes(Dnode) 18: **for** a ∈ A do 19: C ← get valid thresholds(Dnode, a) 20: V ← randomly sample k valid thresholds(C) 21: **for** v ∈ V do 22: save threshold statistics(node, Dnode, a, v) 23: **end** **for** 24: *scores* ← compute split scores(node) 25: select optimal split(node, *scores*) 26: **end for** 27: **end if** 28: Dleft, Dright ← split on selected threshold(node, Dnode) 29: node.left = DareTrain(Dleft, d + 1; drmax, k) 30: node.right = DareTrain(Dright, d + 1; drmax, k) 31: **end if** 32: **return** node 33: **end procedure** Algorithm\n**Algorithm: Deleting a training instance from a DaRE tree, (Brophy and Lowd, 2021). Require**: start at the root node. **Input**: node, data point to remove z, depth d, random node depth drmax, threshold candidate size k. **Output**: retrained subtree rooted at node. 1: **procedure** DareUnlearn(node, z, d; drmax, k) 2: update instance counts |Dnode|, |Dnode,1| 3: **if** node is a LeafNode **then** 4: remove z from leaf-instance pointers(node, z) 5: recompute leaf value(node) 6: remove z from database and return 7: **else** 8: update decision node statistics(node, z) 9: **if** node is a RandomNode **then** 10: **if** node.selectedT hreshold is invalid **then** 11: Dnode ← get data from the set of leaf instances(node) \\ {z} 12: **if** node.selectedAttribute(a) is not constant **then** 13: v ← resample threshold ∈ [amin, amax) 14: Dnode,`, Dnode,r ← split on new threshold(node, Dnode, a, v) 15: node.` ← DareTrain(Dnode,`, d + 1; drmax, k) 16: node.r ← DareTrain(Dnode,r, d + 1; drmax, k) 17: **else** 18: node ← DareTrain(Dnode, d; drmax, k) 19: **end if** 20: remove z from database and return 21: **end if** 22: **else** 23: **if** ∃ invalid attributes or thresholds **then** 24: Dnode ← get data from the set of leaf instances(node) \\ {z} 25: resample invalid attributes and thresholds(node, Dnode) 26: **end if** 27: scores ← recompute split scores(node) 28: a, v ← select optimal split(node, scores) 29: **if** optimal split has changed **then** 30: Dnode.left, Dnode.right ← split on new threshold(node, Dnode, a, v) 31: node.left ← DareTrain(Dnode.left, d + 1; drmax, k) 32: node.right ← DareTrain(Dnode.right, d + 1; drmax, k) 33: remove z from database and return 34: **end if** 35: **end if** 36: **if** xa ≤ v **then** 37: DareUnlearn(node.left, z, d + 1; drmax, k) 38: **else** 39: DareUnlearn(node.right, z, d + 1; drmax, k) 40: **end if** 41: **end if** 42: **end procedure** Efficiency\nThe level of random nodes in a DaRE RF, drmax, is an efficiency parameter, with larger values entailing faster unlearning at the cost of predictive performance. DaRE RFs with random nodes have worse performance than the standard random forest. The number of valid thresholds to consider, k, is another efficiency parameter. Reducing k will increase efficiency, however predictive performance suffers Approximate Unlearning (certified unlearning) Approximate unlearning approaches attempt to address these cost related constraints. In lieu of retraining, these strategies: perform computationally less costly actions on the final weights, modify the architecture or filter the outputs. Essentially we relax the exact unlearning problem to give us a probability or a certainty with which we can say whether or not a sample was in the training set or not.\nTo know more about one of the approximate unlearning methods known as Selective Synaptic Dampening check out:\nSSD paper summary ","date":1704240000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1704240000,"objectID":"a721e669c4b3356cd19063393c3c8c74","permalink":"https://vlgiitr.github.io/posts/machine_unlearning/","publishdate":"2024-01-03T00:00:00Z","relpermalink":"/posts/machine_unlearning/","section":"posts","summary":"Widely used machine learning algorithms are able to learn from new data using batch or online training methods but are incapable of efficiently adapting to data removal. Why do we need data removal though you might think.","tags":null,"title":"Machine Unlearning","type":"posts"},{"authors":["Aayan Yadav"],"categories":null,"content":"Over the years neuroscience has inspired many quantum leaps in Artificial Intelligence. One such remarkable development inspired by the visual ventral system of the brain is Disentangled Variational Autoencoders.\nSo first things first -\nWhat are Autoencoders? In a real-world scenario, fewer dimensions may be required to capture the information stored in a particular data point than already present. This is due to the inherent structure of the data.\nAs shown above, in the first image data points are truly random, there is no structure to data so all three x, y, and z coordinates are necessary to represent data. While in the second image, data is restricted to a spiral, there is some structure to data so that it could be represented by just two variables.\nAutoencoder uses neural networks to provide an unsupervised approach to deal with data.\nData is run through a neural network and map it into a lower dimension called the latent dimension. Then that information can be decoded using a decoder. If we increase the dimensions of the latent space we would get a more detailed image but the number of dimensions required for a considerably clear reconstruction might be very less as compared to the original dimensionality .It could also be used for applications like image segmentation, denoising and neural inpainting.\nHow does it work? Basically we compress the information into latent variables using non linear activation function and then run it through the decoder with the aim of recreating the input data by using just the information stored in latent variables. We calculate the reconstruction loss by comparing the output with input then try to minimize this loss by changing the parameters.\nVariational Autoencoders We have a rough idea of autoencoders by now, so the next question which is arises is what are Variatonal Autoencoders(VAEs) and how are they different ?\nIn VAEs unlike traditional autoencoders the input is mapped to a distribution from which data is sampled and fed into the decoder.\nGiven input data $x$ and latent variable $z$ , encoder tries to learn the posterior distribution $p(z|x)$.\nThis posterior is intractable so VAEs use variational inference to approximate it\nVariational Inference : We choose a family of distribution and then fit it to the input data by changing the parameters. This helps us learn a good approximation to intractable distribution.\nBut how do we know if we have a good approximation of the posterior ? The metric we use to determine how close the approximated distribution is to the required posterior is the Kullback-Liebler Divergence.\n$$ \\hat{q}(z)=\\underset{q\\sim Q}{\\operatorname{argmax}} KL(q(z)||p(z|x))\n$$\nHere q(z) is the approximated distribution and Q is the family of distributions of which q is a member.\nOne visible problem with this is that we dont know p(z|x), so we cant calculate KL divergence directly. To deal with this we convert this into optimization problem. We will skip the maths here and directly jump to the results.\n$$ KL(q(z)||p(z|x))=-ELBO(q)+p(x) $$\nHere ELBO is the something called the Evidence Lower Bound. It is the only term dependent on q. So we have to just maximize ELBO to minimize KL divergence and subsequently find good approximation of the posterior distributaion.\nThe Reparameterization trick: If one pays close attention its difficult to not notice an obvious hurdle in this model. We cant run gradient through sampling operations. So how do we train this model ? This is where the Reparameterization trick comes to rescue!\nWe rewrite z as : $z=\\mu +\\sigma \\bigodot \\epsilon$ .\n$\\bigodot$ here represents the elementwise product of matrices or the Hadamard product\n$\\mu$ — Mean of the distribution\n$\\sigma$ —Standard Deviation\n$\\epsilon \\sim N(0,1)$\nThis reparametrization splits the latent representation into deterministic and stochastic parts. Here $\\mu$ and $\\sigma$ are the deterministic quantities that we train by using gradient descent, while $\\epsilon$\nrepresents the stochastic component, introducing randomness and preventing a direct one-to-one mapping of the data.\nWhat do we mean by ‘disentangling’? Neural networks and the information stored in it is often treated a blackbox with no real way to map which artificial neuron contains what information. Infact there is an entire field of AI called Explainable AI (XAI) dedicated to deal with this problem. One significant reason why it\u0026rsquo;s difficult to comprehend and map this information is that artificial neurons don\u0026rsquo;t store information in an organized and compartmentalized form as we perceive it. It wouldn\u0026rsquo;t be inaccurate to state that knowledge is rather \u0026ldquo;entangled.”\nDisentangling refers to making sure that all neurons in latent space learn something different and uncorellated about training data. change in a single latent unit It helps us to compartmentalise and organise information enabling crucial applications like knowledge transfer and zero-shot learning\nKnowledge Transfer : It is using information learnt in one context to learn new things faster.\nZero-shot learning : It is the use of learnt information to draw inference about unseen data.\nAbility to learn uncorrelated underlying factors in an un supervised setting has far reaching implications. It gives the model the ability to recombine the old information in a novel scenario and extrapolate it to make inference just like humans. It also causes model to learn about basic visual concepts like ‘objectness’. This is crucial in order to make machines that think like humans.\nHow is disentangling executed ? Disentangling is inspired by Visual Ventral System of Brain. We translate the biological constraints to mathematical constraints to apply similar pressures.\nExposure to data with transform continuities : Ventral visual system of infants learn from continously transforming data. Response properties of neurons in the inferior temporal cortex arise through a Hebbian learning algorithm that relies on the fact that nearest neighbours of a particular object in pixel space are the transforms of of the same object.\nThe image above clearly demonstrates that sparse data point do not provide enough information for an unsupervised model to identify where the data manifold should lie.\nThus it is important that the factors of variation of observed data are densely sampled from their respective distributions.\nRedundancy reduction and encouraging statiscal independence : Deep unsupervised model is encouraged to perform redundancy reduction and learn statistically independent factors from continuous data in order to learn basic visual concepts similar to humans\nRedundancy :Difference between maximum entropy a channel can transmit, and the entropy of messages actually transmitted.\nRedundancy reduction is facilitated through learning statistically independent factors\nThis mathematically translates to the following constrained optimisation problem\n$$ \\mathcal{L}(\\theta,\\phi;x)= \\mathbb{E}{q{\\phi}(z|x)}[logp_{\\theta}(x|z)] -\\beta D_{KL}(q_{\\phi}(z|x)||p(z)) $$\nHere we need to maximize $\\mathcal{L}(\\theta,\\phi;x)$ ;\nwhere, $x$ is observed data ;$z \\in \\R^{n}$ are the latent factors; $\\beta \\ge 0$ is the inverse tempreature or regularisation coefficient\nWe generally set the disentangled prior to be isotropic gaussian i.e. $p(z)=\\mathcal{N}(0,I)$\nRedundancy reduction is enforced by constraining the capacity of latent information channel $z$ while preserving enough information to enable reconstruction.\nIsotropic nature of Gaussian puts implicit independence pressure on the latent posterior.\nVarying $\\beta$ changes degree of applied learning pressure during training.\n$\\beta$ =0 ⇒ Standard Maximum Likelihood Learning\n$\\beta$ =1 ⇒ Bayes Solution\nExample: The above image shows difference in latent representations of disentangled and entangled learning on same dataset of 2D shapes.\nIn fig A i.e. disentangled learning with $\\beta$ =4 ; latent factor z5, z7, z4, z9, z2 encode information about position in Y, position in X, scale, cos and sin rotational coordinates respectively. While orther latent factors learn uninformative Gaussian distribution.\nClearly in fig B i.e. the entangled case, there is no such seperation of factors and it is impossible to know what factor encodes what.\nConclusion: The development of Artificial General Intelligence(AGI) i.e. giving machines abililty to learn, think and reason out like humans has been a scientific fantasy for a long time now. Learning of basic visual concepts like objectness, ability to accelerate learning using prior knowledge and ability to infer in a unseen scenario by combining past knowledge are essential qualities for realisation of this goal. Development of unsupervised learning models like disentangled VAEs is a key step in this direction. Its application in Reinforcement learning scenarios is also very promising.\nReferences : Disentangled VAE\u0026rsquo;s (DeepMind 2016) Original VAE paper (2013) ","date":1698192000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1698192000,"objectID":"6ddff78db52b3d356ba2f8593a1123d7","permalink":"https://vlgiitr.github.io/posts/vae/","publishdate":"2023-10-25T00:00:00Z","relpermalink":"/posts/vae/","section":"posts","summary":"Over the years neuroscience has inspired many quantum leaps in Artificial Intelligence. One such remarkable development inspired by the visual ventral system of the brain is Disentangled Variational Autoencoders.\nSo first things first -","tags":null,"title":"Dismantling Disentanglement in VAEs","type":"posts"},{"authors":null,"categories":null,"content":"Over the years neuroscience has inspired many quantum leaps in Artificial Intelligence. One such remarkable development inspired by the visual ventral system of the brain is Disentangled Variational Autoencoders.\n","date":1693872000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1693872000,"objectID":"4c77918f2a7ad1c0bda41c11348a70bd","permalink":"https://vlgiitr.github.io/blogs/distangled_vae/","publishdate":"2023-09-05T00:00:00Z","relpermalink":"/blogs/distangled_vae/","section":"blogs","summary":"Over the years neuroscience has inspired many quantum leaps in Artificial Intelligence. One such remarkable development inspired by the visual ventral system of the brain is Disentangled Variational Autoencoders.","tags":null,"title":"Dismantling Disentanglement in VAEs","type":"blogs"},{"authors":null,"categories":null,"content":"I decided to ask a certain popular language model how to build an explosive, from everday items (for no particular reason), but it didn\u0026rsquo;t give me a plausible answer. What is happening here?\n","date":1693094400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1693094400,"objectID":"8911ccd859890c1989d463df47f4c1c9","permalink":"https://vlgiitr.github.io/blogs/aligned_llm/","publishdate":"2023-08-27T00:00:00Z","relpermalink":"/blogs/aligned_llm/","section":"blogs","summary":"I decided to ask a certain popular language model how to build an explosive, from everday items (for no particular reason), but it didn\u0026rsquo;t give me a plausible answer. What is happening here?","tags":null,"title":"Adversarial Attacks on Aligned Language Models","type":"blogs"},{"authors":["Ishan Garg"],"categories":null,"content":"I decided to ask a certain popular language model how to build an explosive, from everday items (for no particular reason), but it didn\u0026rsquo;t give me a plausible answer. What is happening here?\nHave you ever wondered how would publicly available LLMs respond if asked how to destroy the humanity or how to build an atom bomb?? Well ,turns out they don’t respond to such questions.So what is the reason. Turns out, most LLMs today are trained on text scraped over internet and contains a lot of objectionable content, and in order to prevent the model from answering such questions “aligning” has been done.\nSo in this blog let us try to understand a new approach based on a recently published paper “Universal and Transferable Adversarial Attacks on Aligned Language Models” to bypass this alignment and produce virtually nay objectionable content.Let’s begin!!\nIt is widely known that making small changes to the input of a machine learning model can significantly change its output. Similar techniques have been used against Large Language Models (LLMs), which are powerful language models. Researchers have discovered certain “jailbreaks”, which are cleverly designed input prompts that can make LLMs generate inappropriate or objectionable content. However, unlike traditional adversarial examples that are generated automatically, these jailbreaks are created through human creativity and ingenuity, involving a lot of manual effort to trick the models into producing undesirable results.\nSo what are we doing different that allows us to produce objectionable content?? In layman terms our attack appends a adversarial suffix to the query that attempts to induce negative behavior i.e. produce “dangerous” content. Our suffix token consists of 3 key elements:-\n1) Initial Affirmative Responses\nOur attack targets the model to begin its response with “Sure, here is (content of query)” in response to a number of prompts eliciting undesirable behavior.\n2) Combined greedy and gradient-based discrete optimization\nWe take advantage of token-level gradients to identify a set of potential single-token replacements. Once we have this set, we evaluate the loss for each candidate in the set. After evaluating them, we select the best substitution that maximizes the likelihood of a successful attack.\n3)Robust multi-prompt and multi-model attacks\nWe want suffixes that work for multiple prompts across multiple prompts. So we use our greedy gradient method to find a string that works for all prompts across different models.\nNow let’s look at the technical aspect of the method.\nAssume a user writes the following prompt.\nTell me how to build a bomb So the LLM behind a chatbot reads it in the following form.\nSystem: You are a chat assistant designed to provide helpful and not harmful responses to user queries. User: Tell me how to build a bomb. Assistant: In a normal case the LLM would not provide a response.But we introduce our suffix in the following form to get a response.\nSystem: You are a chat assistant designed to provide helpful and not harmful responses to user queries. User: Tell me how to build a bomb. ! ! ! ! ! ! ! ! ! ! Assistant : In order to create a universal attack we do not consider changes to the user input but we try to find the text in place of ! so that any LLM for any prompt responds affirmatively to the user prompt.\nNow lets look at the key elements in detail.\nProducing Affirmative Responses\nFirst we need to define a loss function to optimize the suffix. We could use log likelihood but it comes with its own set of problems. Instead we would use the approach specified above. The output of LLM would be like this\nSystem: You are a chat assistant designed to provide helpful and not harmful responses to user queries. User: Tell me how to build a bomb. ! ! ! ! ! ! ! ! ! ! Assistant: Sure, here is how to build a bomb: The intuition of this approach is that if the language model can be put into a “state” where this completion is the most likely response, as opposed to refusing to answer the query, then it likely will continue the completion with precisely the desired objectionable behavior. This can be attributed to the autoregressive nature of the LLMs.\nIn multimodal LLMs specifying the first target token was found to be sufficient but in case of text-only space there is a chance that the suffix could overwrite the entire prompt thus getting a response but not the intended one.\nNow let’s have a look at the optimization problem.\nIt denotes the probability that the next token is xn+1 given previous n tokens .\nWe try to minimize the negative log likelihood of probability of target of sequences from x = n+1 to x = n+H where n is the input size.\nGreedy oordinate Gradient-based Search\nA primary challenge in optimizing is that we have to optimize over a discrete set of inputs.\nHere in the algorithm we use gradients with respect to each token to find a set of promising candidates for replacement at each token position.\nHere `I` is the set of the positions of the adversarial suffix. So in the loop we first try to find the k substitutions having lowest gradients for all the positions.Then we initialize elements for each batch by selecting elements at random from the substitution set and then find the batch for which the loss function is minimum.\nUniversal Multi-prompt and Multi-model attacks\nNow we build upon the above algorithm to optimize the attack for multiple prompts.Unlike in the above algorithm here x represents the prompts by the user. We use multiple prompts and their corresponding losses and define a postfix `p` of length l tokens.Instead of specifying a different subset of modifiable tokens for all the prompts we choose a single postfix and optimize the losses over that. Similar to above approach we first find the top -K substitutions for the first prompt by optimizing over p.We start with only first prompt and increment the prompts only when the postfix yields results on the earlier prompts.\nAfter finding the k substitutions the process is similar to the process in the previous algorithm.To make the adversarial examples transferable, we incorporate loss functions over multiple models.\nResults\nFollowing results were obtained on using the above method\nWe find that combining multiple GCG prompts can further improve ASR on several models. Firstly, we attempt to concatenate three GCG prompts into one and use it as the suffix to all behaviors. The “+ Concatenate” row of Table 2 shows that this longer suffix particularly increases ASR from 47.4% to 79.6% on GPT-3.5 (gpt-3.5-turbo), which is more than 2× higher than using GCG prompts optimized against Vicuna models only.\nThe method proposed raise substantial questions regarding current methods for the alignment of LLMs.\nReferences\nPaper on Universal and Transferable Adversarial Attacks on Aligned Language Models\nPhoto by Joshua Hoehne on Unsplash\n","date":1693094400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1693094400,"objectID":"30c3d31cc8e685251cec8848a9ebfa88","permalink":"https://vlgiitr.github.io/posts/attacks_on_aligned_llms/","publishdate":"2023-08-27T00:00:00Z","relpermalink":"/posts/attacks_on_aligned_llms/","section":"posts","summary":"I decided to ask a certain popular language model how to build an explosive, from everday items (for no particular reason), but it didn\u0026rsquo;t give me a plausible answer. What is happening here?","tags":null,"title":"Adversarial Attacks on Aligned Language Models","type":"posts"},{"authors":["Jitesh Jain, ✱Jiachen Li, ✱MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi"],"categories":null,"content":"","date":1687737600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1687737600,"objectID":"631ff7188399661e31684b82113e4fe1","permalink":"https://vlgiitr.github.io/publication/oneformer/","publishdate":"2023-06-26T00:00:00Z","relpermalink":"/publication/oneformer/","section":"publication","summary":"Universal Image Segmentation is not a new concept. Past attempts to unify image segmentation in the last decades include scene parsing, panoptic segmentation, and, more recently, new panoptic architectures. We propose OneFormer, a universal image segmentation framework that unifies segmentation with a multi-task train-once design.","tags":[],"title":"OneFormer: One Transformer to Rule Universal Image Segmentation","type":"publication"},{"authors":null,"categories":null,"content":"DL Discusions by VLG for the Fall Semester 2022 start on 24th September. Stay Tuned\n","date":1661990400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1661990400,"objectID":"6cf9682c99d834f4d6baba1fca54122a","permalink":"https://vlgiitr.github.io/recents/workshop2021/","publishdate":"2022-09-01T00:00:00Z","relpermalink":"/recents/workshop2021/","section":"recents","summary":"DL Discusions by VLG for the Fall Semester 2022 start on 24th September. Stay Tuned","tags":null,"title":"DL Discussions Fall Semeseter 2022","type":"recents"},{"authors":null,"categories":null,"content":" We conduct discussions every week where we dicuss and recent advancements in the field of Deep Learning. Join our Discord to attend the discussions!\nSee the link below for all the resources\nDiscussions Date Topic 22-01-2022 Neural Rendering 29-01-2022 Multi-Model AI 05-02-2022 Transformers 12-02-2022 AlphaCode 19-02-2022 Cross-breeding Transformers and CNNs Workshops Date Topic Resources 19-03-2022 Transfer Learning Colab Notebook 26-03-2022 Introduction to RL Colab notebook All the resources for this semester are compiled here ","date":1659916800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1659916800,"objectID":"91aaa2763e0ff06f44e0ba447d87f87e","permalink":"https://vlgiitr.github.io/previous_discussions/spring_2022_discussion/","publishdate":"2022-08-08T00:00:00Z","relpermalink":"/previous_discussions/spring_2022_discussion/","section":"previous_discussions","summary":"We conduct discussions every week where we dicuss and recent advancements in the field of Deep Learning. Join our Discord to attend the discussions!\nSee the link below for all the resources","tags":null,"title":"Spring 2022 Discussions","type":"previous_discussions"},{"authors":null,"categories":null,"content":"VLG now organises weekly quizes on out Instagram. Hop in every Wednesday and flex your DL knowledge !\n","date":1659916800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1659916800,"objectID":"f20e46d72b7c8895e959713dc3f8720b","permalink":"https://vlgiitr.github.io/recents/weekly_quiz/","publishdate":"2022-08-08T00:00:00Z","relpermalink":"/recents/weekly_quiz/","section":"recents","summary":"VLG now organises weekly quizes on out Instagram. Hop in every Wednesday and flex your DL knowledge !","tags":null,"title":"Weekly AI quiz on Instagram","type":"recents"},{"authors":["✱Jitesh Jain","✱Yuqian Zhou","Ning Yu","Humphrey Shi"],"categories":null,"content":"","date":1659657600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1659657600,"objectID":"c24b1a324151be7a271731379a57667c","permalink":"https://vlgiitr.github.io/publication/imageimpainting/","publishdate":"2022-08-05T00:00:00Z","relpermalink":"/publication/imageimpainting/","section":"publication","summary":"In this paper, we revisited the core design ideas of stateof-the-art deep inpainting networks. We propose an intuitive and effective inpainting architecture that augments the powerful comodulated StyleGAN2 generator with the high receptiveness ability of FFC to achieve equally good performance on both textures and structures.","tags":[],"title":"Keys to Better Image Inpainting: Structure and Texture Go Hand in Hand","type":"publication"},{"authors":null,"categories":null,"content":"We are starting our forthnightly infographic series on Instagram\n","date":1659312000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1659312000,"objectID":"119f104db14ea1ff2f96320607fbc733","permalink":"https://vlgiitr.github.io/recents/ai_bi_weekly_infographs/","publishdate":"2022-08-01T00:00:00Z","relpermalink":"/recents/ai_bi_weekly_infographs/","section":"recents","summary":"We are starting our forthnightly infographic series on Instagram","tags":null,"title":"DL Infographics on Instagram","type":"recents"},{"authors":null,"categories":null,"content":"Members of VLG Harsh Kumar, \u0026ldquo;Kumar Devesh\u0026rdquo;, \u0026ldquo;Sarthak Gupta\u0026rdquo;, participated in the High Prep Event - \u0026ldquo;Bosch model extraction attack for video classification\u0026rdquo; and grabbed GOLD Medal. Congratulations !!\n","date":1657152000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1657152000,"objectID":"c6e7842ae03418357a7ff3dd234e29a2","permalink":"https://vlgiitr.github.io/recents/inter-iit/","publishdate":"2022-07-07T00:00:00Z","relpermalink":"/recents/inter-iit/","section":"recents","summary":"Members of VLG Harsh Kumar, \u0026ldquo;Kumar Devesh\u0026rdquo;, \u0026ldquo;Sarthak Gupta\u0026rdquo;, participated in the High Prep Event - \u0026ldquo;Bosch model extraction attack for video classification\u0026rdquo; and grabbed GOLD Medal. Congratulations !!","tags":null,"title":"Members of VLG grab Gold Medal in Inter-IIT Tech Meet 10.0","type":"recents"},{"authors":["Divyam Goel","Kunal Pratap Singh","Jonghyun Choi"],"categories":null,"content":"","date":1654041600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1654041600,"objectID":"8407779be3b12cf241a879c7e0197394","permalink":"https://vlgiitr.github.io/publication/metacontrolforembodiedinstructionfollowing/","publishdate":"2022-06-01T00:00:00Z","relpermalink":"/publication/metacontrolforembodiedinstructionfollowing/","section":"publication","summary":"Embodied Instruction Following (EIF) is a challenging problem requiring an agent to infer a sequence of actions to achieve a goal environment state from complex language and visual inputs. We propose a generalised Language Guided Meta-Controller (LMC) for better language grounding in the large action space of the embodied agent. We additionally propose an auxiliary reasoning loss to improve the ‘conceptual grounding’ of the agent. Our empirical validation shows that our approach outperforms strong baselines on the Execution from Dialogue History (EDH) benchmark from the TEACh benchmark.","tags":[],"title":"Language Guided Meta-Control for Embodied Instruction Following","type":"publication"},{"authors":["Divyam Goel","Raksha Sharma"],"categories":null,"content":"","date":1654041600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1654041600,"objectID":"a34cff7bb634d4c3840fb35f19cef75f","permalink":"https://vlgiitr.github.io/publication/levragingdependencygrammer/","publishdate":"2022-06-01T00:00:00Z","relpermalink":"/publication/levragingdependencygrammer/","section":"publication","summary":"In this paper, we address the problem of offensive language detection on Twitter, while also detecting the type and the target of the offence. We propose a novel approach called SyLSTM, which integrates syntactic features in the form of the dependency parse tree of a sentence and semantic features in the form of word embeddings into a deep learning architecture using a Graph Convolutional Network. Results show that the proposed approach significantly outperforms the state-of-the-art BERT model with orders of magnitude fewer number of parameters.","tags":[],"title":"Leveraging Dependency Grammar for Fine-Grained Offensive Language Detection using Graph Convolutional Networks","type":"publication"},{"authors":null,"categories":null,"content":"On hearing the term \u0026ldquo;password-cracking,\u0026rdquo; many will think this post will be about how to guess someone\u0026rsquo;s password or somewhat similar, but the reality is not always so satisfying.\n","date":1653868800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1653868800,"objectID":"f9e9c6ea53e9eb799401705cbe08a7e2","permalink":"https://vlgiitr.github.io/blogs/password_cracking/","publishdate":"2022-05-30T00:00:00Z","relpermalink":"/blogs/password_cracking/","section":"blogs","summary":"On hearing the term \u0026ldquo;password-cracking,\u0026rdquo; many will think this post will be about how to guess someone\u0026rsquo;s password or somewhat similar, but the reality is not always so satisfying.","tags":null,"title":"Password Cracking","type":"blogs"},{"authors":["Rohan Garg"],"categories":null,"content":"On hearing the term \u0026ldquo;password-cracking,\u0026rdquo; many will think this post will be about how to guess someone\u0026rsquo;s password or somewhat similar, but the reality is not always so satisfying.\nWhat is Password Cracking In general, whenever anybody types a password on any device or software, passwords don\u0026rsquo;t get stored in the raw format in the database. Instead, raw passwords are first passed through the hashing algorithm, which converts the raw passwords into some particular sequence of letters, numbers, and special characters which looks entirely random for an ordinary being.\nNow there are several password database leaks and breaches all over the world. One such dataset is Rockyou Dataset, which contains about 31 million passwords; this is a widely used dataset because this dataset contains passwords in plain text format without any hashing. Most password cracking algorithms are either trained using this dataset or have used this for dictionary attacks.\nThese algorithms decrypt the hashes of the passwords obtained from other password databases leaks. These algorithms generate passwords, hashes them using the encryption algorithm, and then compare the hash with the hashes present in the database; if the hash matches, Bingo! We got the password corresponding to that hash; otherwise, keep generating and comparing passwords. Password hashes generated by the encryption algorithm are such that they can\u0026rsquo;t be reverted. Passwords, once hashed, can not be converted back into passwords by any algorithm other than brute force attacks over the hash of every possible password.\nHashcat Hashcat is one of the most popular and widely used password crackers. It uses various kinds of attacks for cracking the passwords like:\nDictionary attack: Trying all the passwords present in a list or database. Combinator attack: Trying concatenating words from multiple wordlists. Mask attack: Trying all the characters given in charsets, per position. Hybrid attack: Trying combining wordlist and masks. Association attack: Use a piece of information that could have had an influence on password generation to attack a specific hash. In addition to these, Hashcat enables high-parallelized password cracking and the ability to support a distributed hash cracking system via overlays.\nProbabilistic Context-Free Grammar Context-free grammars have been in the study of natural languages, where they are used to generate strings with a particular structure. Probabilistic context-free grammar is a probabilistic approach to traditional context-free grammar; it incorporates available information about the probability distribution of user passwords. This information is used to generate password patterns in order of decreasing probability. At the same time, these structures can be either password guesses or word-mangling templates that can be filled by dictionary words. Here\u0026rsquo;s a brief overview of how probabilistic context-free grammar is used in password cracking:\nPreprocessing: In this phase, frequencies of specific patterns are measured associated with the password string. In this, the author denotes the alpha string (sequence of alphabet symbols) by L, digit string as D, and special strings(sequence of non-alpha and non-digit symbols) as S. For password \u0026ldquo;$password123\u0026rdquo;, structure of the password would be SLD, base structure would also be similar to structure except that it would also incorporate the length of strings, so base structure would be S¹L⁸D³. The preterminal structure fills in the value of S and D in the base structure, whereas the terminal structure (guess) would fill the value of L in the preterminal structure.\nUsing Probabilistic Grammars: A mathematical form of defining context-free grammar as G = (V, Σ, S, P), where: V is a finite set of variables (or non-terminals), Σ is a finite set of terminals, S is the start variable, and P is a finite set of productions of the form α → β where α is a single variable and β is a string consisting of variables or terminals. Probabilistic context-free grammars have probabilities associated with each production such that for a specific left-hand-side variable, all the associated productions add up to 1. A string derived from the start symbol is called a sentential form. The probability of sentential form is simply the product of the possibilities of the productions used in its derivation. As the production rules don\u0026rsquo;t have any data to rewrite alpha variables to alpha strings, thus sentential forms can be maximally derived up to the terminal digits and special characters with alpha variables. These sentential forms are the pre-terminal structures. The main idea is that preterminal structures define mangling rules that can be directly used in a distributed password cracking trial on passing them to the distributed system to fill in the alpha variables with dictionary words and hash the guesses.\nAssigning pre-terminal structure with probability\nIn order to generate pre-terminal structures in decreasing order of probability, authors used the approach to output all the probable pre-terminal structures, evaluate them on probability, and then sort the results. However, this pre-computation step is not parallelizable with the password cracking step that follows. Now to generate terminal structures from the pre-terminal structure, one approach is to simply fill in all relevant dictionary words for the highest pre-terminal structure and then choose the next highest probable pre-terminal structure. This approach does not further assign probabilities to the dictionary words and does not learn the specific replacement of alpha variables from the training set. This approach is called pre-terminal probability order. Another approach is to assign probabilities to alpha strings in various ways. For instance, it is possible to assign probabilities to words in a dictionary based on how many words of that length appear, observed use of the word, frequency of appearance in language, or knowledge about the target. This approach is called terminal probability order. This approach does assign each terminal structure (password guesses) a well-defined probability.\nFor comparing the performance of probabilistic context-free grammars, the authors used a standard open-source password cracking program, John the Ripper. The authors used a total of six publicly available input dictionaries to use in our tests. Four of them, \u0026ldquo;English_lower\u0026rdquo;, \u0026ldquo;Finnish_lower\u0026rdquo;, \u0026ldquo;Swedish_lower\u0026rdquo; and \u0026ldquo;Common_Passwords\u0026rdquo; were obtained from John the Ripper\u0026rsquo;s public website. Additionally \u0026ldquo;dic-0294\u0026rdquo; input dictionary was obtained from a password-cracking site, and \u0026ldquo;English-wiki\u0026rdquo; input dictionary is based on the English words gathered from www.wiktionary.org.\nNumber of passwords cracked against Myspace list\nPasswords Cracked by the Terminal probability order approach of Probabilistic context-free grammar are the highest. It gave an improvement over John the Ripper from 28% to 129% more passwords cracked given the same number of guesses. Additionally, when we used the preterminal order, we also achieved better results than John the Ripper in all cases but one, though less than what we achieved using terminal probability order.\nPassGAN PassGAN is an example of generative adversarial networks (GANs). GANs are essentially an adversarial framework of multilayer perceptions made up of a generator and discriminator. The generator tries to generate data samples similar to the training data and fool the discriminator. In contrast, the discriminator tries to maximize the probability of assigning the correct label to both the training examples and samples generated by the generator. They both end up playing the minimax game and optimizing the value function V (G, D) for the password distributions.\nGenerative modeling relies on closed-form expressions that generally aren\u0026rsquo;t capable of noisy real data. PassGAN trains a generative deep neural network that takes as input a multi-dimensional random sample of passwords formed in a Gaussian distribution to generate a sample of the desired target distribution.\nThe generator of PassGAN takes input to the reshape node followed by five residual blocks, whereas each block consists of 1D convolutional blocks connected by relu functions, and the final output is the weighted sum of outputs from the conv1D block and the residual identity connection of input. Residual blocks are then followed by a 1D convolutional node which outputs to the softmax node to generate probability distribution in the character set.\nThe discriminator of PassGAN has a very similar architecture to the generator, except that it is in the opposite order as compared to generator. Input after a transpose operation is fed to a 1D convolutional block which is followed by five residual blocks whose architecture is similar to residual blocks used in the case of generator. Output from the final residual block after having a reshape operation is given a linear transformation function which leads to the final output.\nFor the training purpose, 2.5 million passwords were sampled uniformly from the RockYou dataset, length of the passwords was restricted to 10 in order to make training computationally feasible. For testing purposes, additional 2.5 million passwords were sampled exclusive to the training set passwords. To evaluate the trained model, 5 million passwords are generated from the generator network and compared to the test data, about 5.5% (274965) generated passwords were found in the data set, whereas 63110 among them were unique. For calculating the strength of passwords cracked by PassGAN, researchers used zxcbvn. zxcbvn is a low-budget password strength estimator. Its algorithm returns an integer strength bar from 0 - 4, estimating a higher strength with a higher score. Passwords, those PassGAN was able to crack scored 1.59 and landed an average guess per password to be 5.32 x 10⁶.\nGENPass As we have seen, PCFG(Probabilistic context-free grammar), is based on statistical probability. These approaches require a large amount of calculation, which is time-consuming. In PassGAN, neural networks are able to predict more accurate passwords, however, they are not qualified for cross-site attacks as each dataset has its own features.\nGENPass tries to generalize on those leaked passwords and improve the performance in cross-site attacks. GENPass is a multi-source deep learning model that learns from several datasets and ensures the output wordlist can maintain high accuracy for different datasets using adversarial generation. Now before we proceed further, first define what is \u0026ldquo;general\u0026rdquo;.\nDefinition (what is “general”) : Assume a training set T containing m leaked password datasets D1, D2, D3,…,Dm. Model Gt is obtained by training T. Model Gi is obtained by training Di (i ∈ [1, m]). If Model Gt can guess dataset Dk (Dk∉ D1, D2, D3,…,Dm) better than Gi (i ∈ [1, m]), model Gt is believed to be general.\nFor generating passwords, PCFG + LSTM models, also called PL models, comes into the picture. The preprocessing step is performed by PCFG. Passwords are first encoded into a sequence of units. Each unit has a char and a number. A char stands for a sequence of letters (L), digits (D), special chars (S), or an end character (\u0026rsquo;\\n\u0026rsquo;), and the number stands for the length of the sequence. A table is generated when we preprocess the passwords. LSTM is a widely used RNN variant, which generates the probability of the next element based on the context elements. Each LSTM model unit maintains a state Ct at time t, and three sigmoid gates control the data flow in the unit, namely the input gate, the forget gate, and the output gate. The output is calculated as follows:\nLSTM is used to generate passwords. By feeding the LSTM model the preprocessed wordlist and training it, the model can predict the next unit. When a unit is determined, it is transformed back into an alphabetic sequence according to the table generated during the preprocessing step. The LSTM model will output a list of units with their corresponding probabilities, if units are chosen according to the highest weight, then a large number of duplicates will be created in the output wordlist, so the unit is chosen by sampling from discrete distribution. This ensures that higher-weight candidates are chosen with higher probability, while lower ones can still be chosen after a number of guesses. This procedure is called weight choosing.\nPL model is suitable for only one dataset, not for several datasets simultaneously. Different datasets have different underlying principles and lengths, whereas simply mixing datasets would make it difficult for the model to learn the general principles. To solve this multi-source training problem, GENPass comes into the picture.\nPrediction of Model n: For all the different datasets, we train a different PL model, thus, the model can output the result with its own principle.\nWeight Choosing: It is assumed that all the PL model have the same probability, so the output from each model are combined, the combined list will be the input of the weight choosing process, and the final output will be chosen by sampling from discrete distribution.\nClassifier and Discriminator: The classifier is a CNN classifier trained by raw passwords without preprocessing from different datasets. Given a password, the classifier can tell which dataset the password most likely comes from. Through a softmax layer, the output will be a list of numbers with a sum of one. Discriminator takes the classifier\u0026rsquo;s output and accepts those passwords that have a consistent probability of appearance in different datasets so that the output passwords can be \u0026ldquo;general\u0026rdquo;.\nIf C is too large, the generated unit will be discarded; otherwise, it will be accepted. In the model threshold value of C is set to 0.2.\nEvaluation: To evaluate the PL model, it is trained with Myspace and phpBB password datasets. After each training session, the model generated a new wordlist. GENPass is also trained on the same Myspace, and phpBB password datasets and wordlist are generated. The authors trained the PL model with a single mixture of two wordlists and compared the result with the GENPass model.\nHere it is clear that the GENPass model outperforms all other models. Using raw LSTM without any preprocessing performs the worst. Using PL to learn Myspace alone performs second best, which proves Myspace is a good dataset. Simply mixing two datasets does not improve the matching rate.\nReferences Hashcat official page\nAdversarial Password Cracking\nGENPass: A Multi-Source Deep Learning Model for Password Guessing\nPassword Cracking Using Probabilistic Context-Free Grammars\n","date":1653868800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1653868800,"objectID":"a53b61b5e5ee0a962021f77b03b4cfc9","permalink":"https://vlgiitr.github.io/posts/password_cracking/","publishdate":"2022-05-30T00:00:00Z","relpermalink":"/posts/password_cracking/","section":"posts","summary":"On hearing the term \u0026ldquo;password-cracking,\u0026rdquo; many will think this post will be about how to guess someone\u0026rsquo;s password or somewhat similar, but the reality is not always so satisfying.\nWhat is Password Cracking In general, whenever anybody types a password on any device or software, passwords don\u0026rsquo;t get stored in the raw format in the database.","tags":null,"title":"Password Cracking","type":"posts"},{"authors":["Divyam Goel","Ramansh Grover","Fatemeh H Fard"],"categories":null,"content":"","date":1651363200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1651363200,"objectID":"39e179e6d1b623b95005998b623f970d","permalink":"https://vlgiitr.github.io/publication/crossmodeltransfenlp/","publishdate":"2022-05-01T00:00:00Z","relpermalink":"/publication/crossmodeltransfenlp/","section":"publication","summary":"Pre-trained neural Language Models (PTLM), such as CodeBERT, are recently used in software engineering as models pre-trained on large source code corpora. Although adapters are known to facilitate adapting to many downstream tasks compared to fine-tuning the model that require retraining all of the models' parameters -- which owes to the adapters' plug and play nature and being parameter efficient -- their usage in software engineering is not explored. ","tags":[],"title":"On The Cross-Modal Transfer from Natural Language to Code through Adapter Modules","type":"publication"},{"authors":["Jun Chen","Aniket Agarwal","Sherif Abdelkarim","Deyao Zhu","Mohamed Elhoseiny"],"categories":null,"content":"","date":1648512000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1648512000,"objectID":"bf8e05feabe0b5a3c4072cf875472761","permalink":"https://vlgiitr.github.io/publication/reitransformer/","publishdate":"2022-03-29T00:00:00Z","relpermalink":"/publication/reitransformer/","section":"publication","summary":"The visual relationship recognition (VRR) task aims at understanding the pairwise visual relationships between interacting objects in an image. This paper shows that modeling an effective message-passing flow through an attention mechanism can be critical to tackling the compositionality and long-tail challenges in VRR. The method, called RelTransformer, represents each image as a fully-connected scene graph and restructures the whole scene into the relation-triplet and global-scene contexts.","tags":[],"title":"RelTransformer: A Transformer-Based Long-Tail Visual Relationship Recognition","type":"publication"},{"authors":["Jitesh Jain","Anukriti Singh","Nikita Orlov","Zilong Huang","Jiachen Li","Steven Walton","Humphrey Shi"],"categories":null,"content":"","date":1640217600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1640217600,"objectID":"d3bcb7977b5992bae924fb8f44555844","permalink":"https://vlgiitr.github.io/publication/semask/","publishdate":"2021-12-23T00:00:00Z","relpermalink":"/publication/semask/","section":"publication","summary":"Finetuning a pretrained backbone in the encoder part of an image transformer network has been the traditional approach for the semantic segmentation task. However, such an approach leaves out the semantic context that an image provides during the encoding stage. This paper argues that incorporating semantic information of the image into pretrained hierarchical transformer-based backbones while finetuning improves the performance considerably. To achieve this, we propose SeMask, a simple and effective framework that incorporates semantic information into the encoder with the help of a semantic attention operation. In addition, we use a lightweight semantic decoder during training to provide supervision to the intermediate semantic prior maps at every stage. Our experiments demonstrate that incorporating semantic priors enhances the performance of the established hierarchical encoders with a slight increase in the number of FLOPs. We provide empirical proof by integrating SeMask into each variant of the Swin-Transformer as our encoder paired with different decoders. Our framework achieves a new state-of-the-art of 58.22% mIoU on the ADE20K dataset and improvements of over 3% in the mIoU metric on the Cityscapes dataset.","tags":[],"title":"SeMask: Semantically Masked Transformers for Semantic Segmentation","type":"publication"},{"authors":null,"categories":null,"content":" We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning. Join our MS Team to attend the discussions! Team Code : z1q54os\nBasic Discussions We discuss a few fundamental concepts on Wednesdays.\nDate Topic Resources 18-08-2021 Introduction to GANs Slides 25-08-2021 VAE Slides 01-09-2021 Sequence Modelling Slides 08-09-2021 Transformers and Attention Slides 29-09-2021 Reinforcement Learning Slides Workshops Date Topic Resources 13-10-2021 Transfer Learning Jupyter Notebook 20-10-2021 VAE Jupyter Notebook Advanced Discussions We discuss the latest papers published in top tier conferences on Saturdays.\nDate Paper 1 Link Paper 2 Link 18-08-2021 Per-Pixel Classification is Not All You Need for Semantic Segmentation Paper Distilling the Knowledge in a Neural Network Paper 25-08-2021 Large Scale Image Completion via Co-Modulated Generative Adversarial Networks Paper 09-11-2021 Diverse Part Discovery: Occluded Person Re-Identification With Part-Aware Transformer Paper 02-10-2021 Towards Compact CNNs via Collaborative Compression Paper Is Space-Time Attention All You Need for Video Understanding? Paper 09-10-2021 Rethinking Attention with Performers Paper Reformer: The Efficient Transformer Paper 16-10-2021 Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks Paper ","date":1634688000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1634688000,"objectID":"0649e2f185e12f0fbe6a203e2da9ce7c","permalink":"https://vlgiitr.github.io/previous_discussions/aut_2021_discussion/","publishdate":"2021-10-20T00:00:00Z","relpermalink":"/previous_discussions/aut_2021_discussion/","section":"previous_discussions","summary":"We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning. Join our MS Team to attend the discussions! Team Code : z1q54os","tags":null,"title":"Autumn 2021 Discussions","type":"previous_discussions"},{"authors":["Aniket Agarwal*","Sherif Abdelkarim*","Panos Achlioptas","Jun Chen","Jiaji Huang","Boyang Li","Kenneth Church","Mohamed Elhoseiny"],"categories":null,"content":"","date":1632528000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1632528000,"objectID":"6f8ed6552a7a0921dfef78e5cc16fe37","permalink":"https://vlgiitr.github.io/publication/longtailvisualrelationshipwithlargevocab/","publishdate":"2021-09-25T00:00:00Z","relpermalink":"/publication/longtailvisualrelationshipwithlargevocab/","section":"publication","summary":"Several approaches have been proposed in recent literature to alleviate the long-tail problem, mainly in object classification tasks. In this paper, we make the first large-scale study concerning the task of Long-Tail Visual Relationship Recognition (LTVRR). LTVRR aims at improving the learning of structured visual relationships that come from the long-tail (e.g., rabbit grazing on grass). In this setup, the subject, relation, and object classes each follow a long-tail distribution. To begin our study and make a future benchmark for the community, we introduce two LTVRR-related benchmarks, dubbed VG8K-LT and GQA-LT, built upon the widely used Visual Genome and GQA datasets. We use these benchmarks to study the performance of several state-of-the-art long-tail models on the LTVRR setup. Lastly, we propose a visiolinguistic hubless (VilHub) loss and a Mixup augmentation technique adapted to LTVRR setup, dubbed as RelMix. Both VilHub and RelMix can be easily integrated on top of existing models and despite being simple, our results show that they can remarkably improve the performance, especially on tail classes.","tags":[],"title":"Exploring Long tail Visual Relationship Recognition with Large Vocabulary","type":"publication"},{"authors":null,"categories":null,"content":" We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning.\nBasic Discussions We discuss a few fundamental concepts on Wednesdays.\nDate Topic Resources 14-04-2021 Linear Algebra Slides 21-04-2021 Probability and Stats Slides 05-05-2021 Neural Networks and CNNs Slides-A Slides B Advanced Discussions We discuss the latest papers published in top tier conferences on Saturdays.\nDate Paper 1 Link Paper 2 Link 20-02-2021 GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields https://arxiv.org/abs/2011.12100 Swapping Autoencoder for Deep Image Manipulation https://taesung.me/SwappingAutoencoder/ 27-02-2021 ActionBytes: Learning from Trimmed Videos to Localize Actions https://openaccess.thecvf.com/content_CVPR_2020/papers/Jain_ActionBytes_Learning_From_Trimmed_Videos_to_Localize_Actions_CVPR_2020_paper.pdf The Importance of Modeling Data Missingness in Algorithmic Fairness: A Causal Perspective https://arxiv.org/abs/2012.11448 13-03-2021 Ultra-Data-Efficient GAN Training: Drawing A Lottery Ticket First, Then Training It Toughly https://arxiv.org/abs/2103.00397v1 An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale https://arxiv.org/abs/2010.11929 24-04-2021 PREDICT \u0026amp; CLUSTER: Unsupervised Skeleton Based Action Recognition https://arxiv.org/abs/1911.12409 Dynamic Convolutions: Exploiting Spatial Sparsity for Faster Inference https://arxiv.org/abs/1912.03203 01-05-2021 SCOUT: Self-aware Discriminant Counterfactual Explanations https://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_SCOUT_Self-Aware_Discriminant_Counterfactual_Explanations_CVPR_2020_paper.pdf High-performance brain-to-text communication via imagined handwriting https://www.biorxiv.org/content/10.1101/2020.07.01.183384v1.full.pdf ","date":1629331200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1629331200,"objectID":"2025241f434a6b3d90f9c0d78169b010","permalink":"https://vlgiitr.github.io/previous_discussions/spring_2021_discussion/","publishdate":"2021-08-19T00:00:00Z","relpermalink":"/previous_discussions/spring_2021_discussion/","section":"previous_discussions","summary":"We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning.\nBasic Discussions We discuss a few fundamental concepts on Wednesdays.","tags":null,"title":"Spring 2021 Discussions","type":"previous_discussions"},{"authors":null,"categories":null,"content":"The new blog on the topic Riding the Noisy Research Track by Jitesh Jain is now published on Medium. In this blog, one of our members shares his experience and learnings in research, and covers some essential guidelines for a beginner in this field. Do give the blog a read here.\n","date":1627430400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1627430400,"objectID":"46d1675f184201ea9ecf94f6607766b1","permalink":"https://vlgiitr.github.io/recents/noisy_research_blog/","publishdate":"2021-07-28T00:00:00Z","relpermalink":"/recents/noisy_research_blog/","section":"recents","summary":"The new blog on the topic Riding the Noisy Research Track by Jitesh Jain is now published on Medium. In this blog, one of our members shares his experience and learnings in research, and covers some essential guidelines for a beginner in this field.","tags":null,"title":"New Blog on Learnings during research published on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"It\u0026rsquo;s pretty common to get fascinated by the idea of research. But sometimes we lose intrest midway through it. Ride this noisy track with one of our undergrad reasercher !\n","date":1627430400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1627430400,"objectID":"091b82496896fdd3c510362ea4e0d807","permalink":"https://vlgiitr.github.io/blogs/noisy_research_track/","publishdate":"2021-07-28T00:00:00Z","relpermalink":"/blogs/noisy_research_track/","section":"blogs","summary":"It\u0026rsquo;s pretty common to get fascinated by the idea of research. But sometimes we lose intrest midway through it. Ride this noisy track with one of our undergrad reasercher !","tags":null,"title":"Riding the Noisy Research Track","type":"blogs"},{"authors":["Jitesh Jain"],"categories":null,"content":"Alright, people! This article will share my experience and learnings during the last eight months as an undergrad researcher. For those reading one of my blogs for the first time, I am a CSE undergrad (about to enter 3rd year) and am working as a Research Intern at SHI Lab @ UO and Picsart.\nNow, before we dive in, let me clear it out that I have probably had a more intense research experience for better or worse than some of the other undergrad people at IITR. So, don’t make any judgments about research based on the content in this post 😛. So, we got that sorted out. Let’s begin!\nMy research life is going soooooo good :3\nDisclaimer: If it wasn’t clear by now, this isn’t a post where I will tell you how to get a research intern. You will find a lot of those with a single google search.\nThe Promising Start Now, when you are just starting with research and working with a pro researcher, you can’t wait to take up a research project and get into the mix. That was the case with me, no exceptions, each cell in my body was excited to start “my own” research project(s).\nSo, the journey begins!\nIt all begins with a literature survey, reading tons of papers, getting to know the dreaded so-perfect SOTA (state-of-the-art), coming up with the so-called unique innovations, and then finding another paper with the innovations already implemented. That’s right. It isn’t as easy to find a new problem to work on amidst the immense ocean of existing works. It took me almost one month at the start of both my interns to find a project to work on!\nSo, don’t get all stressed and messed up if you don’t have anything to work on at the start; paper reading is the most underrated skill without which you can’t excel in research. Feed your mind with ideas from the others and create a new one with a mix-match 😉. That’s what research is: building on existing works and improvising! That brings me to the next section about building.\nLet’s Build the Concept. Great! You have a concept to work on, but can you build it?\nHere, now you will have two ways to proceed:\nFirst: Take up an open-source implementation (I am so grateful that these exist 🙈) and build on it Second: Start the implementation from scratch all by yourself. Here, DON’T let the craze of learning sway you towards the second option. You learn much more building upon an existing implementation: the code architecture, good practices, and most importantly, how to put modules from different places together in a bug-free (more on this later) manner. On the other hand, doing all the implementation yourself will be a waste of time, effort, and, well, inherent smartness 😛. (To be clear, I didn’t make this mistake, but the thought did cross my mind once ¯\\_(ツ)_/¯)\nWhy is this important? Well, as the research community expands, more and more people are open-sourcing their code, and most of the time, you will use existing works as your baseline, and it is essential to get used to their setting and working, which brings me to the next section.\nThis paper looks good. I often used to ask myself: What’s the significance of the many research conferences held around the globe? The papers are probably months old, outdated, and already read by the targeted audience by the time the conference is organized. Sure, you get to talk to the authors, participate in challenges \u0026amp; workshops, and what else?\nIt turns out conferences are like a quality-assurance mark (well, most of the time). If a work has been accepted to a reputed conference, it’s an excellent decision to take that as your baseline. If not, then well, you better do an autopsy on the claimed results.\nAs a fun fact, there was a work about 3–4 months back unpublished, just floating out on the arxiv. I took that as a baseline like any good kid would do. So, the work was fluked, the implementation and most probably the idea as well. And even more interesting, the primary author knew about the issues. That’s when I invented the rule:\nNever trust a paper blindly, ever!\nBtw, I also decided this. It took a long time coming 🙈\nUnderstand the DATA! DL models are data-hungry!\nIt’s a common saying in the DL community, and for a good reason. Before diving into the network structure code, starting the experiments, and all that goes on during the sprint, it’s nice to prioritize understanding the dataset and checking its correctness. You might think most papers used this dataset, so let’s load the contents and start! That’s right, and at this point, we need to verify if the pre-processing steps work as expected and we didn’t tweak the input in the wrong way unknowingly.\nIt happened with me once, where I took the data processing script from one of the works and used it straight away only to find a week later that the inputs were wrong and not synced to my task 😕. So, respect the data inputs. Without a correct input, you get senseless outputs!\nWhat is this?! Why didn’t I check the data inputs ?🤦\nDO it YOUR Way! As an undergrad researcher, you generally have an advisor who is either a grad student or the prof himself might be the advisor. Either way, don’t expect your advisor to help you in the implementation process. During the meetings, often, they may advise you to change the pipeline in a certain way. Here, understand the logic and implement it in your way. Don’t just follow orders from the advisor. They have probably never looked at the codebase and so have no idea about its state. You created the codebase, so make changes to the code in your way.\nDon’t worry if you make big fat bug blunders. That’s how you learn the art of debugging, and with time, the efficiency increases. So,\nTake the what, why, a little how from the advisor and answer the complete how yourself!\nHaving your recipe is more fun😛\nThe Crisis. Cool, so you have a concept, a credible baseline, and starting implementation. You are ready to start the experiments and then publish your work with a bang. Life is awesome!\nPrank Alert, you just got research bombed 🙅. If your first set of experiments work well, then you are the luckiest person on the planet! Generally, The Crisis encompasses a few phases:\nOh, this doesn’t work; let’s try that idea instead! One thing didn’t work. You tried changing a few hyperparameters, a few modules, the code is a total mess, and then you throw the current idea out and try a completely new one. That makes sense, but remember that when you are trying different things, be patient, don’t change more than one setting in any of your experiments. Track the effects of the changes and then make a decision.\nPS: I have switched ideas a few times, so it isn’t a big deal, I suppose 🙈.\nThe Never-Ending Trials. Your new idea doesn’t work either. You are already into 3–4 months of experiments. Here is where most people realize that research isn’t the right choice for them, so they opt out, I suppose.\nAt this time, one starts to develop something called the Imposter Syndrome (btw, check this blog by Dale Markowitz about Imposter Syndrome in Software Engineering) if they didn’t earlier already. So here’s what I did to fight that:\nTalking to a few senior researchers about their experience helped a lot. Focusing on the learning and not on producing a paper. At least saying that to myself helped me keep the stress and anxiety to a low level. 😛 Experience is the primary thing I am here for. The undergrad years are bonus years for research. Btw, I am still in this phase, so yeah, way to go!\nWelcome to the world of research!\nThe Sorted Phase.\nThis is the phase where the experiments finally work, the paper rolls out, and it is time to move onto newer projects. I hope to experience it someday 🤞.\nI’m Bored :( Because I am a relatively young researcher and also a little impatient (at least that’s what I feel), I often have those moments when I want to scrap the ongoing project and take up something I find more interesting at that moment. There is no fixed step for this phase, and you might be better off sticking to the current project or starting a new one. It all depends on the project and scope, to be honest.\nResearch Finally Pays off :) I found many seniors telling me back in my first year that undergrad research generally doesn’t earn as much money as the other development-centric activities in India. This is true to an extent, but eventually, you will get a research intern where a development intern package won’t bug you anymore 😛. If you are wondering about me, SHI Lab is an unpaid one, and Picsart pays me.\nDeep Learning Research is a lot of Engineering. I knew that implementing things in deep learning research isn’t an easy job. If you are one of those people who think that all that happens in research is thinking, and implementation is the simple part, then you are in for a bumpy ride!\nIndeed, implementations in software development fields and DL research aren’t too different. In the end, all you have to do in both cases is develop/create the most optimal, efficient algorithm and write a script for executing that. We spend most of the time debugging the code, creating diagrams for the pipeline, and running toy experiments crucial to the research process. It’s a good habit to write the code and debug it in your mind as you go along, creating imagining the expected functionalities of the modules. If you are lucky and alert, you may find a big fluke in the pipeline while writing the code itself \\O/.\nSo, it’s not a bad idea to put your developer/engineer hat on at that time. 😛\nWell, research is more engineering than what I used to think 🙈\nTLDR? Alright, that turned out long. Let me provide you with a TLDR here:\nCarry out literature surveys before starting a project. Choose the baseline carefully and go with an existing implementation if available. Spend some time understanding the dataset. Don’t fully listen to advisors about how to implement a pipeline 😛 Fight the crisis phase gallantly :) Your engineering and research shall pay off! Conclusion Right, so that’s it, guys! I tried to share a few learnings from my short research experience so far. In my opinion, what makes research different from other fields is the enormous need for perseverance while working. We have no idea if the problem even has a solution or not. We are wanderers, traversing the planes of experiments to find a positive answer, losing our way in the middle, getting back on track again, which makes it exciting tbh!\nIf you have any questions, opinions (appreciation included :P), or want to discuss your experience (non-research included), contact me!\nI wrote another blog about research. Wow!\nYou can learn more about me on my webpage.\n","date":1627430400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1627430400,"objectID":"87b7b6a6e94930d1207ded2a943f7453","permalink":"https://vlgiitr.github.io/posts/noisy/","publishdate":"2021-07-28T00:00:00Z","relpermalink":"/posts/noisy/","section":"posts","summary":"Alright, people! This article will share my experience and learnings during the last eight months as an undergrad researcher. For those reading one of my blogs for the first time, I am a CSE undergrad (about to enter 3rd year) and am working as a Research Intern at SHI Lab @ UO and Picsart.","tags":null,"title":"Riding the Noisy Research Track","type":"posts"},{"authors":null,"categories":null,"content":"The new blog on the topic What doing research as an undergrad can teach you. by Ayush Mangal is now published on Medium. The Blog talks about the writer\u0026rsquo;s experience as a undergrad researcher with the hope that this article might motivate more people to give research a shot even if they don’t want to become a scientist and remove the stigma surrounding research from their mind. Do give the blog a read here.\n","date":1626998400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1626998400,"objectID":"552e93aff4c91019b728ee54938fc8ed","permalink":"https://vlgiitr.github.io/recents/research_deku_blog/","publishdate":"2021-07-23T00:00:00Z","relpermalink":"/recents/research_deku_blog/","section":"recents","summary":"The new blog on the topic What doing research as an undergrad can teach you. by Ayush Mangal is now published on Medium. The Blog talks about the writer\u0026rsquo;s experience as a undergrad researcher with the hope that this article might motivate more people to give research a shot even if they don’t want to become a scientist and remove the stigma surrounding research from their mind.","tags":null,"title":"New Blog on Researching as an Undergrad Published on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"What is research but a blind date with knowledge ? Explore more about researching as an undergrad with our senpai deku!\n","date":1624406400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1624406400,"objectID":"7273b70207985aaea9d3f21579878f2a","permalink":"https://vlgiitr.github.io/blogs/research_deku/","publishdate":"2021-06-23T00:00:00Z","relpermalink":"/blogs/research_deku/","section":"blogs","summary":"What is research but a blind date with knowledge ? Explore more about researching as an undergrad with our senpai deku!","tags":null,"title":"What doing research as an undergrad can teach you.","type":"blogs"},{"authors":["Ayush Mangal"],"categories":null,"content":"\nAlthough my coaching teachers used to say now and then ki\n“itna deep me jaake kya karega topic me, research karni he kya tereko isme?”\nIf someone told me in my first few months at college, that I’ll be spending most of my time pursuing research as an undergrad, I would have probably laughed it off. And honestly, it was accidental for most of the part, a random conversation with a senior and some fateful string of events made me consider research as a direction that I would like to explore and then it was just one thing after the other, and it’s not as if I am considering research as a life-career, it’s more of a just for fun kind of activity.\nAlso, I am not a full-fledged researcher, I am just an undergrad lol, with only a few small research experiences ( you can find more about it here). Most of them happened under the supervision of highly experienced people, and it was more of me doing stuff for their research and not a research of my own per se, and so I shouldn’t be writing this blog, but there are two reasons I am doing so :\nI probably won’t continue to pursue research as a career in future, so I thought I should write down the feeling, the lessons learnt while they are still fresh, just to come back to them after some years and maybe get inspired or stuff. It’s been way too long since my last blog post, and I wanted to write about something ( the real reason lol) So, I strongly believe, that everyone should try to get some research experience in their undergrad, even if you don’t want to pursue academic research, especially if you don’t want to pursue it as a career. But you might find this statement contradictory and ask, what would research teach me? If I don’t want to pursue it as a career, why is this for me?\nWell, that’s what I will talk about in this blog —\nBringing Order to Chaos What is research but a blind date with knowledge?\n— Will Harvey\nThe most obvious reason for getting research experience that comes to my mind is that it teaches you how to approach an abstract, unstructured problem and gradually bring order to all the chaos, structure stuff up, and devise a solution to a seemingly unsolvable problem.\nUsually, research problem statements are poorly specified, and there is no right or wrong answer to them, no one knows how to solve them accurately, and you get to work on the frontiers of human knowledge, and although it’s doubtful that you’ll be able to make the next great discovery that will lead humankind to salvation, you might be able to take a small small part in the conversation of experts of the world and maybe help shift the conversation from one point to another ( or you can get effectively remain unknown and practically dead for the community, and well, there’s a graveyard for all such dead ideas — arxiv.org 😛)\nThis ability is precious in whatever career you want to pursue, problems in the real world are like research problems too, they don’t have one correct answer ( maybe they don’t have any answer at all ), but you still need to make hypotheses, test it, and then repeat, and that’s precisely what doing research teaches you. I personally feel more confident about tackling any problem in any field after doing some research work, because it has taught me how to approach a problem statement scientifically and break it down into smaller manageable subproblems.\nStanding on the Shoulder of Giants A researcher who believes that he can figure it out all by himself is probably not a very good one. Good researchers respect the opinions of people who came before them, dive deep into their work, what they did, what problems they discovered and how they tackled them. A lot of good research work has been done by taking inspiration from something done by someone lots of years back, which was forgotten over time, and redesigning it into the present. In fact, Issac Newton famously quoted —\nIf I have seen further it is by standing on the shoulders of Giants\n— Issac Newton\nAnd well, I don’t think anyone is vain enough to believe they are smarter than this guy, so if the biggest daddy out there in research says that respect those who came before, you bet we do precisely that.\nAlso, research is a highly collaborative environment. People openly discuss ideas ( albeit, there are some constraints due to the ideas being an intellectual property ), they welcome criticism of their works since they believe that the problem statement they are working on is yet unsolved, they are probably not lucky enough to have it all figured out correctly. And I think this is a bit more prevalent in the US and a bit less in India as of now, but hopefully, that would change in the future.\nWorking in such an environment would pay off in alternate careers, because well, developing a collaborative spirit is quintessential to success. You get comfortable with putting your “ego” aside and look for ideas from everywhere. You become intellectually humble, and begin to realise that one cannot possibly know everything about anything and also begin to respect other people’s work. You stop discarding things as inherently easy because you know that there is probably some really complex research going on even for something seemingly simple. Personally doing research has made me more open to asking for help from others ( sometimes I think I ask for help a little too much lol, but that’s okay ) and also appreciate the inherent complexity of stuff around us.\nThe art of storytelling The human species thinks in metaphors and learns through stories\n— Mary Catherine Bateson\nA critical part of life as a researcher is making other people believe that your work is actually worth considering in the larger conversation going about that problem. And what better than storytelling to make a large group of people believe in something ( take religion, for example, it’s mostly storytelling used to unite a people under a common idealogy ).\nEven while writing a research paper, a lot of effort goes into making stuff sound coherent and in line with the overall “story” you are putting forward. You conduct experiments that further strengthen your story’s validity. You review other people’s version of the story and try to add some quirks of your storytelling into it. A PhD thesis is basically a story of your beliefs, and a PhD defence puts that story to the test, hoping that you can make others believe it.\nAgain this art of storytelling is a vital skill in any career. Say you want to be an entrepreneur, I can’t imagine someone being successful as an entrepreneur without storytelling. You need to make people believe in your story for making them work with you, and you need customers to believe in your story to buy services/products from you. Political parties need people to believe in their story if they want to come to power. The human population has grown to such a level, that to unite these many people, what works the best is making them believe in a common story.\nAlso, as an added bonus, as a researcher, you have to give a lot of presentations ( since you need to present your research work to multiple peoples, at various conferences, guest lectures, reading groups etc. etc. ). So you also get to learn the crucial skill of making great presentations quickly and also a lot of time you have to present complex research work to people who don’t necessarily have the same level of expertise in that area, so you also learn how to simplify complicated stuff and explain it in a simple manner, and that my dear reader, is genius.\nDo or Do not. There is no try. Research teaches a man to admit he is wrong and to be proud of the fact that he does so\n— H.E. Stocher\nThis one is probably the most critical thing research has perhaps taught me personally. Rejection becomes a daily thing for people doing research and failure your only true companion. You work your ass off on a research paper, spending countless hours, doing tedious experiments, and submit it for peer review, and then get it rejected by the whims of a moody reviewer. Also, especially as an undergraduate, research opportunities are sparsely available. You get rejected a lot while applying for research positions, hell most of the people don\u0026rsquo;t even bother replying to undergrads lol.\nAlso doing research requires you to be quite persistent and intrinsically motivated. It’s quite common to feel lost in the process of finding the answer to your question, and help is not readily available, cause well, its a research problem, nobody really knows/cares about it other than you and probably a few more people. You need to stay motivated and keep trying and trying until you finally overcome it or decide to pivot around the issue ( knowing when to pivot and change your hypothesis when stuff isn’t working is also a crucial skill in itself).\nWhen failure becomes a daily thing for you, you start getting used to it; you become used to give it your all, knowing that you might fail. This ability is vital in any career, and it makes you fearless and bold, and more willing to take risks. I personally feel that after getting rejected so many times in my early applications in finding research, I have become quite shameless about failing and have become used to doing whatever it takes to get shit done.\nSome other stuff… Here are some other points that come to my mind, but I don’t want to drag the post too long, so I’ll cover them quickly.\nIt makes you a faster reader and makes you able to gather knowledge pretty quickly, which is a bit obvious since you spend most of the time reading research papers and seeing other people’s work. It makes you more creative since you have to figure out novel solutions to problems day in and out. You can’t just repeat a previous idea, you’ve got to think something new, hell, you might be trying to solve something that no one else did before, you obviously have to be highly creative. It teaches you to respect the small details and how to follow rigorously meticulous experimental procedures. I think doing research has made me better at organising stuff and following the proper procedure of things even if it’s boring. You become better at expressing your ideas, concisely and clearly. Research papers usually have an upper bound on the length of pages, and thus you get used to explaining stuff in as little words as possible ( although you do tend to start using more jargons, but ah well there is no free lunch ) You get used to breaking down a problem to its core and doing things from first principles ( like a physicist ), an approach used by highly successful people like Elon Musk to reimagine the world. While I might not continue pursuing research as a career, I am grateful for whatever research experience I got in the last year. It fundamentally changed the way I approach problems. I hope this article might motivate more people to give it a shot even if they don’t want to become a scientist and remove the stigma surrounding research from their mind ( cause well, even I thought research was for weirdos with a grey beard when I started xD ).\nAs usual, feel free to hit me up if you have any feedback for my writing, I would highly appreciate that! Also hey, let’s talk anyway! Its been a long year, and I am sure you have a lot of interesting stuff to share :P\nUntil then….Stay safe!\n","date":1624406400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1624406400,"objectID":"2b110d6ecbdca3a3a1828b3c0a9de27e","permalink":"https://vlgiitr.github.io/posts/deku/","publishdate":"2021-06-23T00:00:00Z","relpermalink":"/posts/deku/","section":"posts","summary":"Although my coaching teachers used to say now and then ki\n“itna deep me jaake kya karega topic me, research karni he kya tereko isme?”\nIf someone told me in my first few months at college, that I’ll be spending most of my time pursuing research as an undergrad, I would have probably laughed it off.","tags":null,"title":"What doing research as an undergrad can teach you.","type":"posts"},{"authors":null,"categories":null,"content":"The new blog on the topic Optimizer: diving deep into Neural Networks by Rohan Garg is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter. Do give the blog a read here.\n","date":1623888000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1623888000,"objectID":"ea04407d67161f5f767111fba0e349d6","permalink":"https://vlgiitr.github.io/recents/optimizer_blog/","publishdate":"2021-06-17T00:00:00Z","relpermalink":"/recents/optimizer_blog/","section":"recents","summary":"The new blog on the topic Optimizer: diving deep into Neural Networks by Rohan Garg is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter.","tags":null,"title":"New Blog on Optimizers Published on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"Waiting for the neural network to train is always annoying, make sure you use the right optimizers !\n","date":1623888000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1623888000,"objectID":"3b15946390c07fd20c868493efdc0d68","permalink":"https://vlgiitr.github.io/blogs/optimizer/","publishdate":"2021-06-17T00:00:00Z","relpermalink":"/blogs/optimizer/","section":"blogs","summary":"Waiting for the neural network to train is always annoying, make sure you use the right optimizers !","tags":null,"title":"Optimizer: diving deep into Neural Networks","type":"blogs"},{"authors":null,"categories":null,"content":"Waiting for the neural network to train is always annoying, make sure you use the right optimizers !\n","date":1623888000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1623888000,"objectID":"e8e32805e539542cdd15ea9ef4ae0115","permalink":"https://vlgiitr.github.io/posts/optim/","publishdate":"2021-06-17T00:00:00Z","relpermalink":"/posts/optim/","section":"posts","summary":"Waiting for the neural network to train is always annoying, make sure you use the right optimizers !","tags":null,"title":"Optimizer: diving deep into Neural Networks","type":"posts"},{"authors":null,"categories":null,"content":" We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning.\nBasic Discussions We discuss a few fundamental concepts on Wednesdays.\nDate Topic Resources 04-08-2020 Linear Algebra Slides 11-08-2020 Probability Slides 19-08-2020 Neural Networks and CNNs Slides 26-08-2020 Basic CNN architectures Slides 02-09-2020 RNNs Slides 09-09-2020 VAE Slides 16-09-2020 GANs Slides 23-09-2020 Embeddings Slides 30-09-2020 Attention and Transformer Slides 07-10-2020 ELMo and BERT Slides 14-10-2020 RL-I: MDPs, Bellman Equations Slides-A Slides-B 11-11-2020 Graph Neural Networks Slides Advanced Discussions We discuss the latest papers published in top tier conferences on Saturdays.\nDate Paper 1 Link Paper 2 Link 01-08-2020 Reinforced active learning for image segmentation https://arxiv.org/abs/2002.06583 Enhanced POET: Open-Ended Reinforcement Learning through Unbounded Invention of Learning Challenges and their Solutions https://arxiv.org/abs/2003.08536 08-08-2020 Towards Recognizing Unseen Categories in Unseen Domains https://arxiv.org/abs/2007.12256 Neural Arithmetic Units https://arxiv.org/abs/2001.05016 29-08-2020 Learning Memory Access Patterns https://arxiv.org/abs/1803.02329 Equalization Loss for Long-Tailed Object Recognition https://arxiv.org/abs/2003.05176 05-09-2020 PnPNet: End-to-End Perception and Prediction with Tracking in the Loop https://arxiv.org/abs/2005.14711 CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features https://arxiv.org/abs/1905.048993 26-09-2020 Adversarial Continual Learning https://arxiv.org/abs/2003.09553 Decentralized Reinforcement Learning: Global Decision-Making via Local Economic Transactions https://arxiv.org/abs/2007.02382 10-10-2020 Implicit Latent Variable Model for Scene-Consistent Motion Forecasting https://arxiv.org/abs/2007.12036 Large Batch Optimization for Deep Learning: Training BERT in 76 minutes https://arxiv.org/abs/1904.00962 31-10-2020 Causal Discovery with Reinforcement Learning https://arxiv.org/abs/1906.04477 What Should Not Be Contrastive in Contrastive Learning https://arxiv.org/abs/2008.05659 07-11-2020 Dual Super-Resolution Learning for Semantic Segmentation https://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_Dual_Super-Resolution_Learning_for_Semantic_Segmentation_CVPR_2020_paper.pdf Neural Architecture Search without Training Learning https://arxiv.org/abs/2006.04647 28-11-2020 Super-Convergence: Very Fast Training of Neural Networks Using Large Learning Rates https://arxiv.org/abs/1708.07120 ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators https://iclr.cc/virtual_2020/poster_r1xMH1BtvB.html ","date":1606608000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1606608000,"objectID":"8903b47e324c8ff344e86be1eee4735b","permalink":"https://vlgiitr.github.io/previous_discussions/aut_2020/","publishdate":"2020-11-29T00:00:00Z","relpermalink":"/previous_discussions/aut_2020/","section":"previous_discussions","summary":"We conduct two discussions every week where we dicuss the basic concepts and recent advancements in the field of Deep Learning.\nBasic Discussions We discuss a few fundamental concepts on Wednesdays.","tags":null,"title":"Autumn 2020 Discussions","type":"previous_discussions"},{"authors":["Ramit Sawhney","Puneet Mathur","Ayush Mangal*","Piyush Khanna","Rajiv Ratn Shah","Roger Zimmermann"],"categories":null,"content":"","date":1602460800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1602460800,"objectID":"fbf224bdc58b204792221a4815c3fdef","permalink":"https://vlgiitr.github.io/publication/mmtfrf/","publishdate":"2020-10-12T00:00:00Z","relpermalink":"/publication/mmtfrf/","section":"publication","summary":"Stock price movement and volatility prediction aim to predict stocks' future trends to help investors make sound investment decisions and model financial risk. Companies' earnings calls are a rich, underexplored source of multimodal information for financial forecasting. However, existing fintech solutions are not optimized towards harnessing the interplay between the multimodal verbal and vocal cues in earnings calls. In this work, we present a multi-task solution that utilizes domain specialized textual features and audio attentive alignment for predictive financial risk and price modeling. Our method advances existing solutions in two aspects: 1) tailoring a deep multimodal text-audio attention model, 2) optimizing volatility, and price movement prediction in a multi-task ensemble formulation. Through quantitative and qualitative analyses, we show the effectiveness of our deep multimodal approach.","tags":[],"title":"Multimodal Multi-Task Financial Risk Forecasting","type":"publication"},{"authors":null,"categories":null,"content":"Going deeper and deeper works, but only when we know what exactly to learn. Let’s make sure we learn the right thing.\n","date":1601424000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1601424000,"objectID":"9dc4a9ae921643adf994bd1a4a0d2e96","permalink":"https://vlgiitr.github.io/blogs/metric_learning/","publishdate":"2020-09-30T00:00:00Z","relpermalink":"/blogs/metric_learning/","section":"blogs","summary":"Going deeper and deeper works, but only when we know what exactly to learn. Let’s make sure we learn the right thing.","tags":null,"title":"Metric Learning: It’s all about the Distance (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"Going deeper and deeper works, but only when we know what exactly to learn. Let’s make sure we learn the right thing.\n","date":1601424000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1601424000,"objectID":"a46b44795b85af0305a0b76efc7033fe","permalink":"https://vlgiitr.github.io/posts/metric/","publishdate":"2020-09-30T00:00:00Z","relpermalink":"/posts/metric/","section":"posts","summary":"Going deeper and deeper works, but only when we know what exactly to learn. Let’s make sure we learn the right thing.","tags":null,"title":"Metric Learning: It’s all about the Distance (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"The new blog on the topic Metric Learning: It’s all about the Distance by Keerat Kaur Guliani is now published on Medium. The Blog covers the basics of the topic and the common notations followed in subject matter and some of the recent work in this field. Along with this, the blog is supplemented with animations using Manim Engine to further bring forward the main crux of the subject. Do give the blog a read here.\n","date":1601424000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1601424000,"objectID":"99f8476ed2ac0aad2058df85ed6c6382","permalink":"https://vlgiitr.github.io/recents/metric_blog/","publishdate":"2020-09-30T00:00:00Z","relpermalink":"/recents/metric_blog/","section":"recents","summary":"The new blog on the topic Metric Learning: It’s all about the Distance by Keerat Kaur Guliani is now published on Medium. The Blog covers the basics of the topic and the common notations followed in subject matter and some of the recent work in this field.","tags":null,"title":"New Blog on Metric Learning Published on Medium!","type":"recents"},{"authors":["✱Ayush Mangal","✱Jitesh Jain","✱Keerat Kaur Gullani","✱Omkar Bhalero"],"categories":null,"content":"","date":1600473600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1600473600,"objectID":"ee3cd3a2de306e7e4d527f184f472784","permalink":"https://vlgiitr.github.io/publication/deap/","publishdate":"2020-09-19T00:00:00Z","relpermalink":"/publication/deap/","section":"publication","summary":"Recent approaches for learning policies to improve caching, target just one out of the prefetching, admission and eviction processes. In contrast, we propose an end to end pipeline to learn all three policies using machine learning. We also take inspiration from the success of pretraining on large corpora to learn specialized embeddings for the task. We model prefetching as a sequence prediction task based on past misses. Following previous works suggesting that frequency and recency are the two orthogonal fundamental attributes for caching, we use an online reinforcement learning technique to learn the optimal policy distribution between two orthogonal eviction strategies based on them. While previous approaches used the past as an indicator of the future, we instead explicitly model the future frequency and recency in a multitask fashion with prefetching, leveraging the abilities of deep networks to capture futuristic trends and use them for learning eviction and admission. We also model the distribution of the data in an online fashion using Kernel Density Estimation in our approach, to deal with the problem of caching non-stationary data. We present our approach as a ”proof of concept” of learning all three components of cache strategies using machine learning and leave improving practical deployment for future work.","tags":[],"title":"DEAP Cache: Deep Eviction Admission and Prefetching for Cache","type":"publication"},{"authors":null,"categories":null,"content":"In this paper, we propose a DL based approach to tackle the problem of Cache Replacement. This is the first time an approach has tried learning all the three policies: Admission, Prefetching and Eviction. Unlike, previous methods which relied on past statistics for carrying out cache replacement, we predict future statistics (frequency and recency) and then use an online RL-algorithm for eviction.\nCheckout the preprint here.\n","date":1600473600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1600473600,"objectID":"27cac0417b9c5db67abc50e1dd78068c","permalink":"https://vlgiitr.github.io/recents/deap/","publishdate":"2020-09-19T00:00:00Z","relpermalink":"/recents/deap/","section":"recents","summary":"In this paper, we propose a DL based approach to tackle the problem of Cache Replacement. This is the first time an approach has tried learning all the three policies: Admission, Prefetching and Eviction.","tags":null,"title":"Preprint out for the paper: DEAP Cache: Deep Eviction Admission and Prefetching for Cache!","type":"recents"},{"authors":["Aditya Nagori","Raghav Awasthi","Vineet Joshi","Suryatej Reddy Vyalla","Akhil Jarodia","Chandan Gupta","Amogh Gulati","Harsh Bandhey","Keerat Kaur Guliani*","Mehrab Singh Gill","Ponnurangam Kumaraguru","Tavpritesh Sethi"],"categories":null,"content":"","date":1600041600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1600041600,"objectID":"c312a616c6f676b33b0b9272d58740ae","permalink":"https://vlgiitr.github.io/publication/covision/","publishdate":"2020-09-14T00:00:00Z","relpermalink":"/publication/covision/","section":"publication","summary":"COVID-19 pandemic is an enigma with uncertainty caused by multiple biological and health systems factors. Although many models have been developed all around the world, transparent models that allow interacting with the assumptions will become more important as we test various strategies for lockdown, testing and social interventions and enable effective policy decisions. In this paper, we developed a suite of models to guide the development of policies under different scenarios when the lockdown opens. These had been deployed to create an interactive dashboard called COVision which includes the Agent-based Models (ABM) and classical compartmental models (CCM). Our tool allows simulation of scenarios by changing the strength of lockdown, basic reproduction number(R0), asymptomatic spread, testing rate, contact rate, recovery rate, incubation period, leakage in lockdown etc. We optimized ABM and CCMs and evaluated them on multiple error metrics. Out of these models in our suite, ABM was able to capture the data better than CCMs. Our evaluation suggests that ABM models were able to capture the dynamic nature of the epidemic for a longer duration of time while CCMs performed inefficiently. We computed R0 using CCMs which were found to be decreasing with lockdown duration, indicating the effectiveness of policies in different states of India. Models have been deployed on a dashboard hosted at http://covision.tavlab.iiitd.edu.in which allows users to simulate outcomes under different parameters and will allow the policymakers to make informed decisions and efficient monitoring of the covid19 pandemic in India.","tags":[],"title":"Less Wrong COVID-19 Projections With Interactive Assumptions","type":"publication"},{"authors":["Raghav Awasthi","Keerat Kaur Guliani*","Arshita Bhatt","Mehrab Singh Gill","Aditya Nagori","Ponnurangam Kumaraguru","Tavpritesh Sethi"],"categories":null,"content":"","date":1600041600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1600041600,"objectID":"47da2d35e4d58ae5872d354fda5eccf5","permalink":"https://vlgiitr.github.io/publication/vacsim/","publishdate":"2020-09-14T00:00:00Z","relpermalink":"/publication/vacsim/","section":"publication","summary":"A COVID-19 vaccine is our best bet for mitigating the ongoing onslaught of the pandemic. However, vaccine is also expected to be a limited resource. An optimal allocation strategy, especially in countries with access inequities and a temporal separation of hot-spots might be an effective way of halting the disease spread. We approach this problem by proposing a novel pipeline VacSIM that dovetails Actor-Critic using Kronecker-Factored Trust Region (ACKTR) model into a Contextual Bandits approach for optimizing the distribution of COVID-19 vaccine. Whereas the ACKTR model suggests better actions and rewards, Contextual Bandits allow online modifications that may need to be implemented on a day-to-day basis in the real world scenario. We evaluate this framework against a naive allocation approach of distributing vaccine proportional to the incidence of COVID-19 cases in five different States across India and demonstrate up to 100,000 additional lives potentially saved and a five-fold increase in the efficacy of limiting the spread over a period of 30 days through the VacSIM approach. We also propose novel evaluation strategies including a standard compartmental model based projections and a causality preserving evaluation of our model. Finally, we contribute a new Open-AI environment meant for the vaccine distribution scenario, and open-source VacSIM for wide testing and applications across the globe.","tags":[],"title":"VacSIM: Learning Effective Strategies for COVID-19 Vaccine Distribution using Reinforcement Learning","type":"publication"},{"authors":null,"categories":null,"content":"The Basic and Advanced for the new semester begins. We have some really exciting topics lined up to be discussed. With the current pandemic, all the discussions will be online. Just as always, two discussions(Basic and Advanced) will be held every week. We are looking forward to new faces joining in the open discussions. You can also view the discussion schedule and other resources here.\n","date":1596240000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1596240000,"objectID":"c3e3637e5138514b2fae9d94453dfec8","permalink":"https://vlgiitr.github.io/recents/disc/","publishdate":"2020-08-01T00:00:00Z","relpermalink":"/recents/disc/","section":"recents","summary":"The Basic and Advanced for the new semester begins. We have some really exciting topics lined up to be discussed. With the current pandemic, all the discussions will be online. Just as always, two discussions(Basic and Advanced) will be held every week.","tags":null,"title":"The Basic and Advanced Discussions for the Autumn Semester (2020-21) restart!","type":"recents"},{"authors":null,"categories":null,"content":"The core members of the group had a very informative talk and discussion with our alumnus, Shagun Sodhani (Profile). He is currently a research engineer at FAIR and was one of the co-founders of ACM, IIT Roorkee Chapter. He shared experiences of his undergrad years, and also some of the crucial advice to our current members.\n","date":1595721600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1595721600,"objectID":"2a933286e8735d1a17b87174334cedcf","permalink":"https://vlgiitr.github.io/recents/talk_shagun/","publishdate":"2020-07-26T00:00:00Z","relpermalink":"/recents/talk_shagun/","section":"recents","summary":"The core members of the group had a very informative talk and discussion with our alumnus, Shagun Sodhani (Profile). He is currently a research engineer at FAIR and was one of the co-founders of ACM, IIT Roorkee Chapter.","tags":null,"title":"Talk with Alumni: A Talk with Shagun Sodhani","type":"recents"},{"authors":null,"categories":[],"content":"The repo contains summaries of various papers we discuss in our regular discussions and also some other recent papers which we feel have some really exciting contributions for the field.\n","date":1594166400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1594166400,"objectID":"4f5fdbd281e7a59a1525ab571d0a5dcc","permalink":"https://vlgiitr.github.io/project/papers_we_read/","publishdate":"2020-07-08T00:00:00Z","relpermalink":"/project/papers_we_read/","section":"project","summary":"Repo containig summaries we read","tags":[],"title":"Papers We Read","type":"project"},{"authors":null,"categories":null,"content":"The new blog on the topic The Curse of Dimensionality by Jitesh Jain is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter. Along with this, the blog is supplemented with animations using Manim Engine to further bring forward the main crux of the subject. Do give the blog a read here.\n","date":1591833600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591833600,"objectID":"cec0685c558e83df3e60704855b110bc","permalink":"https://vlgiitr.github.io/recents/curse_blog/","publishdate":"2020-06-11T00:00:00Z","relpermalink":"/recents/curse_blog/","section":"recents","summary":"The new blog on the topic The Curse of Dimensionality by Jitesh Jain is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter.","tags":null,"title":"New Blog on Curse of Dimensionality Published on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"Some say breaking this curse is as herculian of a task as breaking the Curse of Medusa. Well, who are we to judge…\n","date":1591833600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591833600,"objectID":"2f0ac668a202af048508bc6cb5314b43","permalink":"https://vlgiitr.github.io/blogs/the-curse-of-dimensionality/","publishdate":"2020-06-11T00:00:00Z","relpermalink":"/blogs/the-curse-of-dimensionality/","section":"blogs","summary":"Some say breaking this curse is as herculian of a task as breaking the Curse of Medusa. Well, who are we to judge…","tags":null,"title":"The Curse of Dimensionality  (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"Some say breaking this curse is as herculian of a task as breaking the Curse of Medusa. Well, who are we to judge…\n","date":1591833600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591833600,"objectID":"1facaa9bed46009d185f8c595068d17f","permalink":"https://vlgiitr.github.io/posts/curse_of_dim/","publishdate":"2020-06-11T00:00:00Z","relpermalink":"/posts/curse_of_dim/","section":"posts","summary":"Some say breaking this curse is as herculian of a task as breaking the Curse of Medusa. Well, who are we to judge…","tags":null,"title":"The Curse of Dimensionality  (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"The new blog on the topic Principal Component Analysis by Ankit Biswas is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter. Along with this, the blog is supplemented with animations using Manim Engine to further bring forward the main crux of the subject. Do give the blog a read here.\n","date":1591142400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591142400,"objectID":"d5161d2ba502dde001c7340ec171a97f","permalink":"https://vlgiitr.github.io/recents/pca_blog/","publishdate":"2020-06-03T00:00:00Z","relpermalink":"/recents/pca_blog/","section":"recents","summary":"The new blog on the topic Principal Component Analysis by Ankit Biswas is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter.","tags":null,"title":"New Blog on Principal Component Analysis on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"Too many dimensions can be bad for your model’s health. Here comes PCA to the rescue.\n","date":1591142400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591142400,"objectID":"018d30ebcb64c85585fde8969cd0206e","permalink":"https://vlgiitr.github.io/blogs/pca/","publishdate":"2020-06-03T00:00:00Z","relpermalink":"/blogs/pca/","section":"blogs","summary":"Too many dimensions can be bad for your model’s health. Here comes PCA to the rescue.","tags":null,"title":"Principal Component Analysis (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"Too many dimensions can be bad for your model’s health. Here comes PCA to the rescue.\n","date":1591142400,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1591142400,"objectID":"ce82fc652284ac787ae8e213d6295034","permalink":"https://vlgiitr.github.io/posts/pca/","publishdate":"2020-06-03T00:00:00Z","relpermalink":"/posts/pca/","section":"posts","summary":"Too many dimensions can be bad for your model’s health. Here comes PCA to the rescue.","tags":null,"title":"Principal Component Analysis (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"The new blog on the topic Singular Value Decomposition by Kaaira Gupta is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter. Along with this, the blog is supplemented with animations using Manim Engine to further bring forward the main crux of the subject. Do give the blog a read here.\n","date":1590451200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1590451200,"objectID":"24227c74d8c2c209a11deb0d83c1cc82","permalink":"https://vlgiitr.github.io/recents/svd_blog/","publishdate":"2020-05-26T00:00:00Z","relpermalink":"/recents/svd_blog/","section":"recents","summary":"The new blog on the topic Singular Value Decomposition by Kaaira Gupta is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter.","tags":null,"title":"New Blog on Singular Value Decomposition on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"Whether you want to compress an image or calculate pseudo-inverse, SVD will always be there for you.\n","date":1590451200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1590451200,"objectID":"abc51495e6b6b54b91cd4cc11b51f49d","permalink":"https://vlgiitr.github.io/blogs/svd/","publishdate":"2020-05-26T00:00:00Z","relpermalink":"/blogs/svd/","section":"blogs","summary":"Whether you want to compress an image or calculate pseudo-inverse, SVD will always be there for you.","tags":null,"title":"Singular Value Decomposition (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"Whether you want to compress an image or calculate pseudo-inverse, SVD will always be there for you.\n","date":1590451200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1590451200,"objectID":"de1bbb6654a29215196ad7c5a0054c0d","permalink":"https://vlgiitr.github.io/posts/svd/","publishdate":"2020-05-26T00:00:00Z","relpermalink":"/posts/svd/","section":"posts","summary":"Whether you want to compress an image or calculate pseudo-inverse, SVD will always be there for you.","tags":null,"title":"Singular Value Decomposition (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"The new survey paper on the topic of Universal Adversarial Perturbations is out now on arxiv. The paper covers the basic terminologies and concepts that build up the subject matter of Adversarial Perturbations and then all the recent literature work to further develop the field. Lastly it also delves into the applications of the same and the future work directions possible in this field.\nCheckout the preprint here.\n","date":1589587200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589587200,"objectID":"5bb221dbd41a8a7ddd6f3352af96fd9a","permalink":"https://vlgiitr.github.io/recents/uap/","publishdate":"2020-05-16T00:00:00Z","relpermalink":"/recents/uap/","section":"recents","summary":"The new survey paper on the topic of Universal Adversarial Perturbations is out now on arxiv. The paper covers the basic terminologies and concepts that build up the subject matter of Adversarial Perturbations and then all the recent literature work to further develop the field.","tags":null,"title":"Preprint out for the paper: Universal Adversarial Perturbations: A Survey!","type":"recents"},{"authors":null,"categories":null,"content":"The new survey paper on the topic of Scene Graphs is out now on arxiv. The paper covers the basic terminologies and concepts that build up the subject matter of Scene Graphs and then all the recent literature work to further develop the field. Lastly it also delves into the applications of the same and the future work directions possible in this field.\nCheckout the preprint here.\n","date":1589587200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589587200,"objectID":"a0d1cc22acc4c7e180c309c2b6e8c181","permalink":"https://vlgiitr.github.io/recents/scene_graph/","publishdate":"2020-05-16T00:00:00Z","relpermalink":"/recents/scene_graph/","section":"recents","summary":"The new survey paper on the topic of Scene Graphs is out now on arxiv. The paper covers the basic terminologies and concepts that build up the subject matter of Scene Graphs and then all the recent literature work to further develop the field.","tags":null,"title":"Preprint out for the paper: Visual Relationship Detection using Scene Graphs: A Survey!","type":"recents"},{"authors":["Ashutosh Chaubey*","Nikhil Agrawal*","Kavya Barnwal*","Keerat K. Guliani*","Pramod Mehta*"],"categories":null,"content":"","date":1589587200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589587200,"objectID":"16a50c58fb043375e836421d367faaa9","permalink":"https://vlgiitr.github.io/publication/uap/","publishdate":"2020-05-16T00:00:00Z","relpermalink":"/publication/uap/","section":"publication","summary":"Over the past decade, Deep Learning has emerged as a useful and efficient tool to solve a wide variety of complex learning problems ranging from image classification to human pose estimation, which is challenging to solve using statistical machine learning algorithms. However, despite their superior performance, deep neural networks are susceptible to adversarial perturbations, which can cause the network's prediction to change without making perceptible changes to the input image, thus creating severe security issues at the time of deployment of such systems. Recent works have shown the existence of Universal Adversarial Perturbations, which, when added to any image in a dataset, misclassifies it when passed through a target model. Such perturbations are more practical to deploy since there is minimal computation done during the actual attack. Several techniques have also been proposed to defend the neural networks against these perturbations. In this paper, we attempt to provide a detailed discussion on the various data-driven and data-independent methods for generating universal perturbations, along with measures to defend against such perturbations. We also cover the applications of such universal perturbations in various deep learning tasks.","tags":[],"title":"Universal Adversarial Perturbations: A Survey","type":"publication"},{"authors":["Aniket Agarwal*","Ayush Mangal*","Vipul Kumar"],"categories":null,"content":"","date":1589587200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589587200,"objectID":"d72de6b13b10d310badf5f37e05e9ed7","permalink":"https://vlgiitr.github.io/publication/scene_graph/","publishdate":"2020-05-16T00:00:00Z","relpermalink":"/publication/scene_graph/","section":"publication","summary":"Understanding a scene by decoding the visual relationships depicted in an image has been a long studied problem. While the recent advances in deep learning and the usage of deep neural networks have achieved near human accuracy on many tasks, there still exists a pretty big gap between human and machine level performance when it comes to various visual relationship detection tasks. Developing on earlier tasks like object recognition, segmentation and captioning which focused on a relatively coarser image understanding, newer tasks have been introduced recently to deal with a finer level of image understanding. A Scene Graph is one such technique to better represent a scene and the various relationships present in it. With its wide number of applications in various tasks like Visual Question Answering, Semantic Image Retrieval, Image Generation, among many others, it has proved to be a useful tool for deeper and better visual relationship understanding. In this paper, we present a detailed survey on the various techniques for scene graph generation, their efficacy to represent visual relationships and how it has been used to solve various downstream tasks. We also attempt to analyze the various future directions in which the field might advance in the future. Being one of the first papers to give a detailed survey on this topic, we also hope to give a succinct introduction to scene graphs, and guide practitioners while developing approaches for their applications.","tags":[],"title":"Visual Relationship Detection using Scene Graphs: A Survey","type":"publication"},{"authors":null,"categories":null,"content":"The new blog on the topic Support Vector Machines by Aaryan Garg is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter. Along with this, the blog is supplemented with animations using Manim Engine to further bring forward the main crux of the subject. Do give the blog a read here.\n","date":1589500800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589500800,"objectID":"925ecf350b28e3af3aa6e33a3749faee","permalink":"https://vlgiitr.github.io/recents/svm_blog/","publishdate":"2020-05-15T00:00:00Z","relpermalink":"/recents/svm_blog/","section":"recents","summary":"The new blog on the topic Support Vector Machines by Aaryan Garg is now published on Medium. The Blog covers the basics of the topic and the mathematical fundamentals followed in subject matter.","tags":null,"title":"New Blog on Support Vector Machines on Medium!","type":"recents"},{"authors":null,"categories":null,"content":"SVM is the Mr. Perfect of Machine Learning Classifiers. It basically wants different examples to be as far as possible.\n","date":1589500800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589500800,"objectID":"4d9fcd4d0ca27681d47cddc6f6effbc1","permalink":"https://vlgiitr.github.io/blogs/svm/","publishdate":"2020-05-15T00:00:00Z","relpermalink":"/blogs/svm/","section":"blogs","summary":"SVM is the Mr. Perfect of Machine Learning Classifiers. It basically wants different examples to be as far as possible.","tags":null,"title":"Support Vector Machine  (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"SVM is the Mr. Perfect of Machine Learning Classifiers. It basically wants different examples to be as far as possible.\n","date":1589500800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1589500800,"objectID":"5d62f2058df3da1155265dd1bb218717","permalink":"https://vlgiitr.github.io/posts/svm/","publishdate":"2020-05-15T00:00:00Z","relpermalink":"/posts/svm/","section":"posts","summary":"SVM is the Mr. Perfect of Machine Learning Classifiers. It basically wants different examples to be as far as possible.","tags":null,"title":"Support Vector Machine  (Medium)","type":"posts"},{"authors":["Arnab Kumar Mondal","Aniket Agarwal","Jose Dolz","Christain Desrosiers"],"categories":null,"content":"","date":1572739200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1572739200,"objectID":"69ad3d23c136903a3f13f5b242bc0259","permalink":"https://vlgiitr.github.io/publication/cycle-gan/","publishdate":"2019-11-03T00:00:00Z","relpermalink":"/publication/cycle-gan/","section":"publication","summary":"In this work, we study the problem of training deep networks for semantic image segmentation using only a fraction of annotated images, which may significantly reduce human annotation efforts. Particularly, we propose a strategy that exploits the unpaired image style transfer capabilities of CycleGAN in semi-supervised segmentation. Unlike recent works using adversarial learning for semi-supervised segmentation, we enforce cycle consistency to learn a bidirectional mapping between unpaired images and segmentation masks. This adds an unsupervised regularization effect that boosts the segmentation performance when annotated data is limited. Experiments on three different public segmentation benchmarks (PASCAL VOC 2012, Cityscapes and ACDC) demonstrate the effectiveness of the proposed method. The proposed model achieves 2-4% of improvement with respect to the baseline and outperforms recent approaches for this task, particularly in low labeled data regime.","tags":[],"title":"Revisiting CycleGAN for Semi-Supervised Segmentation","type":"publication"},{"authors":["Jaynil Jaiswal","Ashutosh Chaubey","Bhimavarapu Sasi Kiran Reddy","Shashank Kashyap","Puneet Kumar","Balasubramanian Raman","Partha Pratim Roy"],"categories":null,"content":"","date":1572652800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1572652800,"objectID":"85afa700c46556d599adceca179e3028","permalink":"https://vlgiitr.github.io/publication/gan-ensemble/","publishdate":"2019-11-02T00:00:00Z","relpermalink":"/publication/gan-ensemble/","section":"publication","summary":"Coming Soon","tags":[],"title":"A Generative Adversarial Network based Ensemble Technique for Automatic Evaluation of Machine Synthesized Speech","type":"publication"},{"authors":["Jogendra Nath Kundu","Maharshi Gor","Dakshit Agrawal","R. Venkatesh Babu"],"categories":null,"content":"","date":1562025600,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1562025600,"objectID":"d25ba51ba7eba728b25708c2fa87f250","permalink":"https://vlgiitr.github.io/publication/gan-tree/","publishdate":"2019-07-02T00:00:00Z","relpermalink":"/publication/gan-tree/","section":"publication","summary":"Despite the remarkable success of generative adversarial networks, their performance seems less impressive for diverse training sets, requiring learning of discontinuous mapping functions. Though multi-mode prior or multigenerator models have been proposed to alleviate this problem, such approaches may fail depending on the empirically chosen initial mode components. In contrast to such bottom-up approaches, we present GAN-Tree, which follows a hierarchical divisive strategy to address such discontinuous multi-modal data. Devoid of any assumption on the number of modes, GAN-Tree utilizes a novel modesplitting algorithm to effectively split the parent mode to semantically cohesive children modes, facilitating unsupervised clustering. Further, it also enables incremental addition of new data modes to an already trained GAN-Tree, by updating only a single branch of the tree structure. As compared to prior approaches, the proposed framework offers a higher degree of flexibility in choosing a large variety of mutually exclusive and exhaustive tree nodes called GANSet. Extensive experiments on synthetic and natural image datasets including ImageNet demonstrate the superiority of GAN-Tree against the prior state-of-the-art.","tags":[],"title":"GAN-Tree: An Incrementally Learned Hierarchical Generative Framework for Multi-Modal Data Distributions","type":"publication"},{"authors":null,"categories":null,"content":"A few problems for the ones looking to test their learning.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"f64bc697741a6a17b70e33c20e550eb2","permalink":"https://vlgiitr.github.io/blogs/advanced_problems/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/advanced_problems/","section":"blogs","summary":"A few problems for the ones looking to test their learning.","tags":null,"title":"Advanced Problems (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"A few problems for the ones looking to test their learning.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"2f9122c69a58a25682261bc575444030","permalink":"https://vlgiitr.github.io/posts/advanced_problems/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/advanced_problems/","section":"posts","summary":"A few problems for the ones looking to test their learning.","tags":null,"title":"Advanced Problems (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the regularization method Dropout.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"d5efa4a2714b178db470ec9ca1a8ed7d","permalink":"https://vlgiitr.github.io/blogs/dropout/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/dropout/","section":"blogs","summary":"Notes on the regularization method Dropout.","tags":null,"title":"Dropout","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the regularization method Dropout.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"4940cedd75d254c83a06fb6e6e428bb4","permalink":"https://vlgiitr.github.io/posts/dropout/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/dropout/","section":"posts","summary":"Notes on the regularization method Dropout.","tags":null,"title":"Dropout","type":"posts"},{"authors":null,"categories":null,"content":"This blog contains a guide to get started with Deep Learning as well as get in-depth knowledge of the field.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"4efec3f954aef0a4c7a21771f11726f0","permalink":"https://vlgiitr.github.io/blogs/dl-guide/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/dl-guide/","section":"blogs","summary":"This blog contains a guide to get started with Deep Learning as well as get in-depth knowledge of the field.","tags":null,"title":"Guide to Deep Learning (Medium)","type":"blogs"},{"authors":null,"categories":null,"content":"This blog contains a guide to get started with Deep Learning as well as get in-depth knowledge of the field.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"3a92248f23ce25c034dd0262e022c66a","permalink":"https://vlgiitr.github.io/posts/dl-guide/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/dl-guide/","section":"posts","summary":"This blog contains a guide to get started with Deep Learning as well as get in-depth knowledge of the field.","tags":null,"title":"Guide to Deep Learning (Medium)","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the the InceptionNet CNN architecture.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"e17dc197f23af92675faf5671e29c5f4","permalink":"https://vlgiitr.github.io/blogs/inception/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/inception/","section":"blogs","summary":"Notes on the the InceptionNet CNN architecture.","tags":null,"title":"InceptionNet","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the the InceptionNet CNN architecture.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"112505929e61151c2b0724ede9f8d71f","permalink":"https://vlgiitr.github.io/posts/inception/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/inception/","section":"posts","summary":"Notes on the the InceptionNet CNN architecture.","tags":null,"title":"InceptionNet","type":"posts"},{"authors":null,"categories":null,"content":"Notes on Linear Algebra from the DL Book.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"458d3ea4f199249fbab4b1ff59f15445","permalink":"https://vlgiitr.github.io/blogs/linalg/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/linalg/","section":"blogs","summary":"Notes on Linear Algebra from the DL Book.","tags":null,"title":"Linear Algebra","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on Linear Algebra from the DL Book.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"480601c90dff75990b989d0876f8aae1","permalink":"https://vlgiitr.github.io/posts/linalg/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/linalg/","section":"posts","summary":"Notes on Linear Algebra from the DL Book.","tags":null,"title":"Linear Algebra","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the Net2Net architecture.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"5d6296a520654ee9e350b490a2b5fdf1","permalink":"https://vlgiitr.github.io/blogs/net2net/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/net2net/","section":"blogs","summary":"Notes on the Net2Net architecture.","tags":null,"title":"Net2Net","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the Net2Net architecture.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"cbbca972b47d0fb3675d3920feeaad25","permalink":"https://vlgiitr.github.io/posts/net2net/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/net2net/","section":"posts","summary":"Notes on the Net2Net architecture.","tags":null,"title":"Net2Net","type":"posts"},{"authors":null,"categories":null,"content":"Notes on Probability and Information Theory from the DL Book.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"12975d78805d07e7f21552c9dba041f2","permalink":"https://vlgiitr.github.io/blogs/probstats/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/probstats/","section":"blogs","summary":"Notes on Probability and Information Theory from the DL Book.","tags":null,"title":"Probability and Information Theory","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on Probability and Information Theory from the DL Book.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"2ce47ceae61286cb75e5f7799d89a1ff","permalink":"https://vlgiitr.github.io/posts/probstats/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/probstats/","section":"posts","summary":"Notes on Probability and Information Theory from the DL Book.","tags":null,"title":"Probability and Information Theory","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the object detection architecture RCNN.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"fded67753bb4e8abb5f7ed0fe1195e41","permalink":"https://vlgiitr.github.io/blogs/rcnn/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/rcnn/","section":"blogs","summary":"Notes on the object detection architecture RCNN.","tags":null,"title":"RCNN","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the object detection architecture RCNN.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"35a9663e24c5e5dbfc2bb948b3f39392","permalink":"https://vlgiitr.github.io/posts/rcnn/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/rcnn/","section":"posts","summary":"Notes on the object detection architecture RCNN.","tags":null,"title":"RCNN","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the CNN Architecture ResNet.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"b69e2fa238ebee884ab57c61a64116c1","permalink":"https://vlgiitr.github.io/blogs/resnet/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/resnet/","section":"blogs","summary":"Notes on the CNN Architecture ResNet.","tags":null,"title":"ResNet","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the CNN Architecture ResNet.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"58b31afdee0db4ff59be87687e7b6ecb","permalink":"https://vlgiitr.github.io/posts/resnet/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/resnet/","section":"posts","summary":"Notes on the CNN Architecture ResNet.","tags":null,"title":"ResNet","type":"posts"},{"authors":null,"categories":null,"content":"Notes on the CNN Architecture VGGNet.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"d7d8aefc79fa098670a3300672e02692","permalink":"https://vlgiitr.github.io/blogs/vgg/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/blogs/vgg/","section":"blogs","summary":"Notes on the CNN Architecture VGGNet.","tags":null,"title":"VGGNet","type":"blogs"},{"authors":null,"categories":null,"content":"Notes on the CNN Architecture VGGNet.\n","date":1558051200,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1558051200,"objectID":"c7e05f9939e56128a5aa04f452e5a1d1","permalink":"https://vlgiitr.github.io/posts/vgg/","publishdate":"2019-05-17T00:00:00Z","relpermalink":"/posts/vgg/","section":"posts","summary":"Notes on the CNN Architecture VGGNet.","tags":null,"title":"VGGNet","type":"posts"},{"authors":["Aarush Gupta","Dakshit Agrawal","Hardik Chauhan","Jose Dolz","Marco Pedersoli"],"categories":null,"content":"","date":1533168000,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":1533168000,"objectID":"bf6aa00b2908fd70c0daaf90170dea54","permalink":"https://vlgiitr.github.io/publication/att/","publishdate":"2018-08-02T00:00:00Z","relpermalink":"/publication/att/","section":"publication","summary":"In this paper we propose a new approach for classifying the global emotion of images containing groups of people. To achieve this task, we consider two different and complementary sources of information: i) a global representation of the entire image (ii) a local representation where only faces are considered. While the global representation of the image is learned with a convolutional neural network (CNN), the local representation is obtained by merging face features through an attention mechanism. The two representations are first learned independently with two separate CNN branches and then fused through concatenation in order to obtain the final group-emotion classifier. For our submission to the EmotiW 2018 group-level emotion recognition challenge, we combine several variations of the proposed model into an ensemble, obtaining a final accuracy of 64.83% on the test set and ranking 4th among all challenge participants.","tags":[],"title":"An Attention Model for Group Level Emotion Recognition","type":"publication"},{"authors":null,"categories":null,"content":"title: \u0026ldquo;Machine Unlearning\u0026rdquo; date: 2024-01-03 math: false diagram: false external_link: \u0026ldquo;https://equinox-rest-93f.notion.site/Machine-Unlearning-84d941dcce56429782c49c9f7280fccd\"\nBlog on machine unlearning.\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"da8514338eac3f2c1791fb1e31f52416","permalink":"https://vlgiitr.github.io/blogs/machine_unlearning/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/blogs/machine_unlearning/","section":"blogs","summary":"title: \u0026ldquo;Machine Unlearning\u0026rdquo; date: 2024-01-03 math: false diagram: false external_link: \u0026ldquo;https://equinox-rest-93f.notion.site/Machine-Unlearning-84d941dcce56429782c49c9f7280fccd\"\nBlog on machine unlearning.","tags":null,"title":"","type":"blogs"},{"authors":null,"categories":null,"content":"VLG goes recruiting for its core members and designers! For the recruitment process this year, we are introducing many ways to become a core team member at VLG. We encourage you to apply through as many ways as you wish.\nRecruitment Test for 1st Year Your familiarity with deep learning concepts will be tested along with a few open-ended questions to analyze your thinking and explaining skills.\nVenue: Room no - 004, APJ Block\nTime: 9th March 7:00 pm to 8:30 pm\nPitch Your Project Submit your ideas in the form of a well drafted proposal\nProposal Submission Deadline: 16th March 11:59pm\nCall For Bloggers If you are a passionate writer and wish to share your knowledge and insights into the world of deep learning with our community, send us your articles. We\u0026rsquo;ll be publishing shortlisted ones through the Medium account of VLG.\nBlog submission deadline: 16th March 11:59pm\nDesign Assignment: We are also looking for designers to join our team. If you have a creative eye and love designing, we would love to hear from you.\nSubmission Deadline: 17th March\nYou can follow us on instagram and find posters for this year here\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"8d41a9ddd0e7e49753dab2e66e906918","permalink":"https://vlgiitr.github.io/recruitments/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/recruitments/","section":"","summary":"VLG goes recruiting for its core members and designers! For the recruitment process this year, we are introducing many ways to become a core team member at VLG. We encourage you to apply through as many ways as you wish.","tags":null,"title":"","type":"page"},{"authors":null,"categories":null,"content":" Open Projects ","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"a2c0d5a69335e11fbd47e53818b3510f","permalink":"https://vlgiitr.github.io/archives/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/archives/","section":"","summary":" Open Projects ","tags":null,"title":"Archives","type":"page"},{"authors":null,"categories":null,"content":"Discussions for this semester will start by the end of September'22. Meanwhile you can look into the previous discussions below.\nPrevious Semester Discussions Spring 2022 Discussions Autumn 2021 Discussions Spring 2021 Discussions Autumn 2020 Discussions ","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"6ded10aea5cf519ccf04b023129fa6a6","permalink":"https://vlgiitr.github.io/discussions/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/discussions/","section":"","summary":"Discussions for this semester will start by the end of September'22. Meanwhile you can look into the previous discussions below.\nPrevious Semester Discussions Spring 2022 Discussions Autumn 2021 Discussions Spring 2021 Discussions Autumn 2020 Discussions ","tags":null,"title":"Discussions","type":"page"},{"authors":null,"categories":null,"content":"You have the four projects below to choose from and you can only pick one. Fill the form with your choice. Join our Open Slack Community or checkout our Instagram Post for more information.\nForm Deadline: 25th May EOD.\nProject Deadline: 22nd June 2023.\nThe project will be verified based on your performance.\nBest Of Luck!\nAI GENERATED IMAGE DETECTION With the rise of generative AI, fake identities can be easily created using sophisticated algorithms. This has led to an increase in identity fraud, as fake identities can be used to gain access to online services and commit fraudulent activities. Thus, the purpose is to contribute to the development of more secure and reliable online services for everyone.\nFull project description: Here\nMULTILINGUAL PRODUCT RECOMMENDER SYSTEM Modelling customer shopping intentions is crucial for e-commerce stores, as it directly impacts user experience and engagement. With this intention, we aim to utilize the ‘Multilingual Shopping Session Dataset’, a dataset consisting of millions of user sessions from six different locales, where the major languages of products are English, German, Japanese, French, Italian, and Spanish.\nFull project description: Here\nSTABLE DIFFUSION - IMAGE TO PROMPT The popularity of text-to-image models has spurned an entire new field of prompt engineering. Part art and part unsettled science, ML practitioners and researchers are rapidly grappling with understanding the relationships between prompts and the images they generate and in this project we aim to reverse the typical direction of a generative text-to-image model: instead of generating an image from a text prompt, can you create a model which can predict the text prompt given a generated image?\nFull project description: Here\nIMAGE MATCHING - RECONSTRUCT 3D SCENES FROM 2D Your best camera may just be the phone in your pocket. You might take a snap of a landmark, then share it with friends. By itself, that photo is two-dimensional and only includes the perspective of your shooting location. Of course, many people may have taken photos of that same landmark. If we were able to combine all of our photos, we may be able to create a more complete, three-dimensional view of any given thing and this is what we aim in this project. Our objective is building a 3D model of a scene given an unstructured collection of images taken around it.\nFull project description: Here\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"cdc04493036c768e740de3f067c89248","permalink":"https://vlgiitr.github.io/open_projects/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/open_projects/","section":"","summary":"You have the four projects below to choose from and you can only pick one. Fill the form with your choice. Join our Open Slack Community or checkout our Instagram Post for more information.","tags":null,"title":"Open Projects","type":"page"},{"authors":null,"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"80992ad071037a77171ba6ee81c21126","permalink":"https://vlgiitr.github.io/alums/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/alums/","section":"","summary":"","tags":null,"title":"Our Alumni","type":"widget_page"},{"authors":null,"categories":null,"content":"Introduction This repo houses summaries for various excitng works in the field of Deep Learning. You can contribute summaries of your own. Check out our contributing guide to start contributing. Happy Reading \u0026amp; Summarizing!\nContents Summaries 2022 2021 2020 2019 2018 2017 2016 Contributing Acknowledgements License Summaries 2022 Human-level play in the game of Diplomacy by combining language models with strategic reasoning [Paper][Review] Meta Fundamental AI Research Diplomacy Team (FAIR), Antin Bakhtun, Noam Brown, Emily Dinan, Science Journal 2022 Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding [Paper][Review] Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, S. Sara Mahdavi, Rapha Gontijo Lopes, Tim Salimans, Jonathan Ho, David J Fleet, Mohammad Norouzi, NIPS 2022 Learning Video Representations from Large Language Models [Paper][Review] Yue Zhao, Ishan Misra, Philipp Krähenbüh, Rohit Girdhar, Facebook AI Research- Meta AI, University of Texas, Austin 2021 GANcraft: Unsupervised 3D Neural Rendering of Minecraft Worlds [Paper][Review] Zekun Hao, Arun Mallya, Serge Belongie, Ming-Yu Liu, ICCV 2021 GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields [Paper][Review] Michael Niemeyer, Andreas Geiger, CVPR 2021 Creative Sketch Genetation [Paper][Review] Songwei Ge, Devi Parikh, Vedanuj Goswami \u0026amp; C. Lawrence Zitnick, ICLR 2021 Binary TTC: A Temporal Geofence for Autonomous Navigation[Paper][Review] Abhishek Badki, Orazio Gallo, Jan Kautz, Pradeep Sen, CVPR 2021 2020 Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild [Paper][Review] Shangzhe Wu, Christian Rupprecht, Andrea Vedaldi, CVPR 2020 You Only Train Once: Loss-conditional training of deep networks [Paper][Review] Alexey Dosovitskiy, Josip Djolonga, ICLR 2020 GrokNet: Unified Computer Vision Model Trunk and Embeddings For Commerce [Paper][Review] Sean Bell, Yiqun Liu, Sami Alsheikh, Yina Tang, Ed Pizzi, M. Henning, Karun Singh, Omkar Parkhi, Fedor Borisyuk, KDD 2020 Semantically multi-modal image synthesis [Paper][Review] Zhen Zhu, Zhiliang Xu, Ansheng You, Xiang Bai, CVPR 2020 Learning to Simulate Dynamic Environments with GameGAN [Paper][Review] Seung Wook Kim, Yuhao Zhou, Jonah Philion, Antonio Torralba, Sanja Fidler, CVPR 2020 Adversarial Policies : Attacking deep reinforcement learning [Paper][Review] Adam Gleave, Michael Dennis, Cody Wild, Neel Kant, Sergey Levine, Stuart Russell, ICLR 2020 Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning [Paper][Review] Jean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec, Pierre H. Richemond, Elena Buchatskaya, Carl Doersch, Bernardo Avila Pires, Zhaohan Daniel Guo, Mohammad Gheshlaghi Azar, Bilal Piot, Koray Kavukcuoglu, Rémi Munos, Michal Valko, CVPR 2020 2019 ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks [Paper][Review] Jiasen Lu, Dhruv Batra, Devi Parikh, Stefan Lee, NIPS 2019 Stand-Alone Self-Attention in Vision Models [Paper][Review] Prajit Ramachandran, Niki Parmar, Ashish Vaswani, Irwan Bello, Anselm Levskaya, Jonathon Shlens, NIPS 2019 Zero-Shot Entity Linking by Reading Entity Descriptions [Paper][Review] Lajanugen Logeswaran , Ming-Wei Chang‡ Kenton Lee , Kristina Toutanova , Jacob Devlin, Honglak Lee ACL-2019 Do you know that Florence is packed with visitors? Evaluating state-of-the-art models of speaker commitment [Paper][Review] Nanjiang Jiang and Marie-Catherine de Marneffe , ACL-2019 Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representations [Paper][Review] Vincent Sitzmann, Michael Zollhofer, Gordon Wetzstein, NIPS-2019 Emotion-Cause Pair Extraction: A New Task to Emotion Analysis in Texts [Paper][Review] Rui Xia, Zixiang Ding, ACL-2019 Putting an End to End-to-End: Gradient-Isolated Learning of Representations [Paper][Review] Sindy Lowe, Peter O\u0026rsquo; Connor, Bastiaan S. Veeling, NIPS-2019 Bridging the Gap between Training and Inference for Neural Machine Translation [Paper][Review] Wen Zhang, Yang Feng, Fandong Meng, Di You, Qun Liu, ACL-2019 Designing and Interpreting Probes with Control Tasks [Paper][Review] John Hewitt, Percy Liang, EMNLP-2019 Specializing Word Embeddings (for Parsing) by Information Bottleneck [Paper][Review] Xiang Lisa Li, Jason Eisner, EMNLP-2019 vGraph: A Generative Model for Joint Community Detection and Node Representational Learning [Paper][Review] Fan-Yun Sun, Meng Qu, Jordan Hoffmann, Chin-Wei Huang, Jian Tang, NIPS-2019 Uniform convergence may be unable to explain generalization in deep learning [Paper][Review] Vaishnavh Nagarajan, J. Zico Kolter, NIPS-2019 SinGAN: Learning a Generative Model from a Single Natural Image [Paper][Review] Tamar Rott Shaham, Tali Dekel, Tomer Michaeli, ICCV-2019 Graph U-Nets [Paper][Review] Hongyang Gao, Shuiwang Ji, ICML-2019 Feature Denoising for Improving Adversarial Robustness [Paper][Review] Cihang Xie, Yuxin Wu, Laurens van der Maaten, Alan Yuille, kaiming He, CVPR-2019 This Looks Like That: Deep Learning for Interpretable Image Recognition [Paper][Review] Chaofan Chen, Oscar Li, Chaofan Tao, Alina Jade Barnett, Jonathan Su, Cynthia Rudin, NIPS-2019 2018 CyCADA: Cycle-Consistent Adversarial Domain Adaptation [Paper][Review] Judy Hoffman, Eric Tzeng, Taesung Park, Jun-Yan Zhu, Phillip Isola, Kate Saenko, Alexei A. Efros, Trevor Darrell, ICML-2018 2017 Unpaired Image-to-Image Translation using Cycle Consistent Adversarial Networks [Paper][Review] Jun-Yan Zhu, Taesung Park, Phillip Isola, Alexei A. Efros, ICCV-2017 Densely Connected Convolutional Networks [Paper][Review] Gao Huang, Zhuang Liu, Laurens van der Maaten, Kilian Q. Weinberger, CVPR-2017 On Calibration of Modern Neural Networks [Paper][Review] Chuan Guo, Geoff Pleiss, Yu Sun, Kilian Q. Weinberger, ICML-2017 2016 Siamese Recurrent Architectures for Learning Sentence Similarity [Paper][Review] Jonas Mueller, Aditya Thyagarajan, AAAI-2016 Contributing We appreciate all contributions to the set of summaries. Please refer to CONTRIBUTING.md for the contributing guideline.\nAcknowledgements papers_we_read is an open source repository that welcomes any contribution and feedback. We wish the collected sets of summaries can help the DL community to start with the practice of reading and understanding research papers which is a potent skill in the research community. Most of our contributors include students enrolled in undergraduate programmes. We are grateful for all the contributions that help improve this collection of summaries.\nLicense This repo is open-sourced under the MIT License.\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"72f310c341e481d13ccfcb85ea3d8e04","permalink":"https://vlgiitr.github.io/summaries/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/summaries/","section":"","summary":"Introduction This repo houses summaries for various excitng works in the field of Deep Learning. You can contribute summaries of your own. Check out our contributing guide to start contributing. Happy Reading \u0026amp; Summarizing!","tags":null,"title":"Papers We Read","type":"page"},{"authors":null,"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"3bf44c81f2a197de01edde90bcd77783","permalink":"https://vlgiitr.github.io/team/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/team/","section":"","summary":"Meet our Team","tags":null,"title":"Team Members","type":"widget_page"},{"authors":null,"categories":null,"content":"","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"b8a927f78b18c2336ffdec19ca890957","permalink":"https://vlgiitr.github.io/recents/recruitments/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/recents/recruitments/","section":"recents","summary":"","tags":null,"title":"VLG goes recruiting","type":"recents"},{"authors":null,"categories":null,"content":" Weekend Talks with researchers is a two day event where researchers from different fields will share their work and experience. Join us on 26 June and 27 June 2021 for the event.\n","date":-62135596800,"expirydate":-62135596800,"kind":"page","lang":"en","lastmod":-62135596800,"objectID":"cb64c48d89213128f9b4a90a909092fe","permalink":"https://vlgiitr.github.io/dre/","publishdate":"0001-01-01T00:00:00Z","relpermalink":"/dre/","section":"","summary":"Weekend Talks with researchers is a two day event where researchers from different fields will share their work and experience. Join us on 26 June and 27 June 2021 for the event.","tags":null,"title":"Weekend Talks with Researchers","type":"widget_page"}]