It's the data, stupid [1] (not the algorithm).
[1] https://en.wikipedia.org/wiki/It%27s_the_economy,_stupid
Twitter's Recommendation Algorithm
941–950 of 1001 posts
Re: Twitter's Recommendation Algorithm
#942Earlier quoted context omitted.
Update: Elon was asked about these in a Twitter Space, he says it's not appropriate and will be removed from the codebase. Additionally, from another Twitter engineer, the Democrat/Republican flags are apparently 10 years old and not important and do not have high feature importance.
Elon seems embarrassed: https://twitter.com/elonmusk/status/1641908130274525187?s=61... It’ll be interesting to see what gets cut. Maybe just the Elon flag, but maybe others too.
Don't let Chesterton's Fence hit your butt on the way out...
Re: Twitter's Recommendation Algorithm
#943Earlier quoted context omitted.
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.
So many questions. How are users tagged D or R? Is that a manual process or automated somehow? What is the effect of these tags? Can I find out if my Twitter account is in one of those buckets?
Likely they are tracking performance verified politician accounts based on registered party affiliation. Why republican should count equal in the evaluation metric to democrat when nunerically there are less republican voters, let alone proportions on Twitter, is another question.
Re: Twitter's Recommendation Algorithm
#944Folks talk about media bias: Twitter popularity is a media bias.
It’s the most lazy journalism to be able to write a “news” article about what Kim, or Don, or Elon’s PR team tweeted.
But, as far as “social” this media is: Twitter is a one-way street. There’s no one actually responding or interacting with Tweets. It’s just a comment section to flame bait.
Maybe we’ll all get lucky and Elon will cause Twitter to go away forever.
Re: Twitter's Recommendation Algorithm
#945Earlier quoted context omitted.
They said that they use it for metrics, so clearly there must be an "elon impression" metric.
Yes, it’s not making anyone see any extra Elon tweets as your comment alleged.
It means they won't ship features that hurt Elon's reach. So in a sense, it is biasing code changes in favour of Elon.
Re: Twitter's Recommendation Algorithm
#946From 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…
As a European I find this very offensive
Re: Twitter's Recommendation Algorithm
#947I think they have a bug here here: https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92...
Code: ( "has_gte_10k_favs", _.getOrElse(EarlybirdFeature, None).exists(_.favCountV2.exists(_ >= 1000))),
Should be: ( "has_gte_10k_favs", _.getOrElse(EarlybirdFeature, None).exists(_.favCountV2.exists(_ >= 10000))),
Re: Twitter's Recommendation Algorithm
#948From 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…
Elon just said in the space "that shouldn't be there. Consider it gone"
[flagged]
Re: Twitter's Recommendation Algorithm
#949Unless a trusted third party, forensically audits Twitter, there is no guarantee the published code corresponds to the actual live code in Production. Also multiple parts are not present as stated in the blog.
This should be seen as a possible snapshot of some code, that might have run, might run in the future, or is possibly running in some parts of the production infrastructure at Twitter.
Re: Twitter's Recommendation Algorithm
#950Great! But nothing is going to change until people realize that the problem is the feedback loop . It's not the recommendation engine itself, it's the fact that there's no way "out" of the feed that the engine produces. It recommends you stuff, you have little choice but to engage with it, and then it trains on that information . This is the problem with most of social media today. It is a very well known problem in…
I feel like the Youtube one is good. You can mark videos and channels as "not interested" and Youtbe really knows me due to my account age and usage... It recommends me unknown videos and I tend to like them but also more mainstream stuff.
It doesn't matter how good they try to make their recommendation system, they will never know you like yourself. For example, when I go to the YouTube home page, there is a list of categories at the top that it's identified for me. I didn't choose these. I can't add or remove them myself if it's wrong. I just have to hope that I watch the "right videos" and it picks up on a new interest that I have.
But I already know what interests I have! I want to have videos about terrariums on my home page now, not in a week when I've watched enough. This is what I mean by recommendation systems not being good enough. They need to give the user more control over what they want to see, because they can never read my mind. Their recommendations will get even better with that information!