ML

this is comprehensive roadmap for machine learning , deep learning and computer vision . By : Dr.Mostafa I. Saad

ML RoadMap

visit the source : https://youtube.com/playlist?list=PLPt2dINI2MIYdFB4H9bTmen9H6sbSzz2_&si=-Tllise7ll5N2oyH

Roadmap for learning Machine Learning (More focus on DL):

General Notes

-- Roadmaps nature

It is not mandatory when you write a roadmap this meaning that i followed this roadmap literally for many reasons like may be there is update in new courses which is more beneficial than the one i have studied and so on .

Recall DL vs Classical Machine Learning

Try to search about their history

MOOCs assignments.

in the online courses are not enough , generally it tells you to complete some functionalities which are very trivial and easy

You should do the whole project by yourself , not just some functionalities

Can't understand something ?

try to search , use GPT , and try to find an abstraction for this topic as / a powerpoint/ from high level universities before getting deeper in that topic

When to start learning ML

generally after the second year of your college as you could finish many course in the faculty related to Maths , Programming and so on .

---

Roadmap as following

• Andrew Ng Coursera Course .

• https://www.coursera.org/learn/machine-learning

• https://medium.com/analytics-vidhya/python-implementation-of-andrew-ngs-machine-learning-course-part-1-6b8dd1c73d80

---

• Kaggle -- Browse Scikit-learn library - use their NN tool (and others if want)

---

Scikit-learn Documentation

https://scikit-learn.org/0.18/_downloads/scikit-learn-docs.pdf

https://blogs.mathworks.com/loren/201...

-- Kaggle: titanic (classification) and Boston housing prices competitions (regression).

• classification project

https://www.kaggle.com/c/titanic

• Regression project

https://www.kaggle.com/c/house-prices...

• Go deeper in Neural Networks -- Why NN?

Try to search about that as NN can do many things without need for the other models

it can act as classifier and regressor and many other reasons -- Deeper understanding

Its very mandatory matter --- Backpropagation

The most important thing you will learn in NN as its great power --- You may listen to 1 or more lecture of these:

 • Lecture 10 - Neur...

!https://www.gstatic.com/youtube/img/watch/yt_favicon.png

---

 • 12a: Neural Nets

!https://www.gstatic.com/youtube/img/watch/yt_favicon.png

---

 • Lecture 4 | Intro...

!https://www.gstatic.com/youtube/img/watch/yt_favicon.png

Great practice :

--- On paper: Compute/Trace an example:

https://mattmazur.com/2015/03/17/a-st...

-- Reimplement NN from scratch --- Multi-class classification (softmax) --- Multi-label classification (sigmoid) --- Regression --- Try AE with your network (AutoEncoder) -- Learn dropout and implement it

-- FYI: NN/ML Arabic playlist:

 • Neural Networks I...

!https://www.gstatic.com/youtube/img/watch/yt_favicon.png

-- Maybe read NN chapter from a book

• Evaluation metrics https://towardsdatascience.com/metric...

-- Later: (when you start computer vision)

https://medium.com/@jonathan_hui/map-...

• AFTER you finish the above, do the following time split strategy

- Next 25% of time for classical Machine Learning models like SVM and so on / projects - 75% for deep learning**

CNN and RNN

---

• 25% of your time for classical ML will be in -- kaggle projects / Kaggle blogs --

https://www.coursera.org/specializati...

-- Feature Engineering for Machine Learning: Principles and Techniques for Data Scientists

75% of your time for deep learning

--

https://www.coursera.org/specializati...

---

 / @deeplearningai

!https://www.gstatic.com/youtube/img/watch/yt_favicon.png

-- [FYI]

https://classroom.udacity.com/courses...

-- Frameworks (TF/Pytorch) - study while learning -- Either learn 2D vision or nlp (later after work learn 2nd) --- Vision key problems: classification, detection, segmentation, pose estimation, object traction, action detection/recognition --- stanford courses ---- Computer Vision:

http://cs231n.stanford.edu/syllabus.html

---- NLP:

http://web.stanford.edu/class/cs224n/

• More on the road -- ML book, e.g. Bishop -- Understand more models, fields, more experience -- Field follow up: E.g. in Vision (CVPR, ICCV, ECCV) -- DL Experience: Paper-to-code skill -- Vision elements (later):

https://classroom.udacity.com/courses...

-- Abstracting skills

• Other ML areas (later) -- Generative models --- PGM, EM (GMM), VI, MCMC --- DL: VAE (variational autoencoder), GAN -(Study GAN)

---

https://medium.com/@jonathan_hui/gan-...

---

https://github.com/nashory/gans-aweso...

-- Reinforcement Learning

---

Updates/Additions

• What if you are also interested in competitive programming? -- ML is time-consuming to be good. In this case, divide your vacation to 50% competitive and 50% for ML