(Nov 2021 Dec 2021) Weekly Sprint Review Calls

Dec 2

Introductions
GPU & Computational resources - looking into options for GPU for ML training
Going into 2022, seeking to partner with organizations to be trusted stewards of the project. (Offer SME, social guidance on racially biased language & context for technical implementation)
Updating stalebot for 60 days to mark an issue as stale, 90 days to automatically close an issue
Going forward changing weekly meeting time to 1pm EST, 11:30pm IST, 10am PST, 6pm GMT, 7pm CET

In order to move forward, we need to figure out computational resources to train the data. This is currently a blocker Preetika has been facing since the start of October so raising to discuss options.
Discussed some difficulties Preetika faced with cleaning the data due to special characters and multiple variations. Francois offered to look into this.
In a future sprint/iteration - we’d like to look into scraping different sites (Twitter, Reddit) respecting privacy & site terms of course.
Leveraging expertise from political scientists, media experts, linguists, social scientists, etc. to guide us in the technical implementation of the nuances faced when identifying biased language. Revisiting the topic of what kind of partners we need on board to do so.
On a logistical plane, can we adjust the stale bot in GitHub to not close issues or adjust the time period in which the issues are closed? I think a lot of the issues are getting closed when they should be open cuz I think we're working on the project at a slower pace than what would be like a regular “work” project. @Upkar (IBM Mentor)
Working on closing out a wiki page for the TakeTwo repo consisting of backstory, FAQs, general roadmap. Hoping to close out this week.