Typically, we expect to be able to run "open source" software ourselves. If you open-source your C compiler, I can compile a C program with it. In a few recent high-profile cases though, companies have "open sourced" ML systems without releasing the model weights. This practice is just like your releasing the builds scripts for your C compiler, but not the compiler itself. While more transparency from social media will be enlightening, calling a release like this (or LLaMA) "open source" feels like equivocation. I'd love to see more full releases, weights included.
Twitter's Recommendation Algorithm
61–70 of 1001 posts
Re: Twitter's Recommendation Algorithm
#62From https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92... ( "author_is_elon", candidate => candidate .getOrElse(AuthorIdFeature, None).contains(candidate.getOrElse(DDGStatsElonFeature, 0L))), ( "author_is_power_user", candidate => candidate .getOrElse(AuthorIdFeature, None) .exists(candidate.getOrElse(DDGStatsVitsFeature, Set.empty[Long]).contains)), ( "author_is_democrat", candidate => candidate .getOrEls…
I read the code snippet before I saw the link and thought you were joking, but yeah, there really is an author_is_elon flag right there in the main branch.
Re: Twitter's Recommendation Algorithm
#63Earlier quoted context omitted.
Twitter has this now. The home page is split into two tabs: "For you", the algorithmic feed, and "Following", the reverse chronological feed of just who you follow.
On my “following” tab (on the phone app) , I’m still getting recommendations for bomb throwers I don’t follow. Am I weird? It’s like an unhinged relative. Not pleasant. Edit: I reversed “for you” and “following” in my original reply.
I sometimes get these as push notifications with my username added to them.
Re: Twitter's Recommendation Algorithm
#64I wonder what determines 'cred' for this part: https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92...
I answered my own question https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92...
"This method reduces the page rank of users who have a low number of followers but a high number of followings."
Re: Twitter's Recommendation Algorithm
#65I'm not opposed to social media feeds having complex recommendation algorithms. I just wish they allowed you to opt in to a reverse chronological feed of only people you follow, like RSS.
Twitter has this now. The home page is split into two tabs: "For you", the algorithmic feed, and "Following", the reverse chronological feed of just who you follow.
Twitter has always had "chronological timeline" behind a confusing "sparkle" button (except for a brief period a few months back where they removed it, or always defaulted to back to algo timeline? and then restored it a week later)
They called it "Latest Tweets" https://web.archive.org/web/20200205092104/https://help.twit...
Re: Twitter's Recommendation Algorithm
#66From https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92... ( "author_is_elon", candidate => candidate .getOrElse(AuthorIdFeature, None).contains(candidate.getOrElse(DDGStatsElonFeature, 0L))), ( "author_is_power_user", candidate => candidate .getOrElse(AuthorIdFeature, None) .exists(candidate.getOrElse(DDGStatsVitsFeature, Set.empty[Long]).contains)), ( "author_is_democrat", candidate => candidate .getOrEls…
The author_is_elon flag doesn't surprise me, but the two political designators are somewhat shocking. I'd sure like to know what changes based on what Twitter knows about your political affiliation.
The repo suggests it's about tracking engagement metrics[0], so Team Red people see more Team Red content and vice versa. Nothing nefarious.
[0]https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92...
Re: Twitter's Recommendation Algorithm
#67Re: Twitter's Recommendation Algorithm
#68From https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92... ( "author_is_elon", candidate => candidate .getOrElse(AuthorIdFeature, None).contains(candidate.getOrElse(DDGStatsElonFeature, 0L))), ( "author_is_power_user", candidate => candidate .getOrElse(AuthorIdFeature, None) .exists(candidate.getOrElse(DDGStatsVitsFeature, Set.empty[Long]).contains)), ( "author_is_democrat", candidate => candidate .getOrEls…
I wonder who's on the "VIT" (Very Important Tweeter) list?
Re: Twitter's Recommendation Algorithm
#69This is pretty limited. I picked a term used in the diagram to see what I could find out about it. But there seems to be next to nothing in the released code about the mentioned "author diversity". No real code or description.
Re: Twitter's Recommendation Algorithm
#70the-algorithm is such a pretentious name for a repo
Eh, it's name-spaced.