Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Twitter's Recommendation Algorithm

blog.twitter.com

Post reply on HN
Print view View original

11–20 of 1001 posts

Re: Twitter's Recommendation Algorithm

#11

Post by tric »

From 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
          .getOrElse(AuthorIdFeature, None)
          .exists(candidate.getOrElse(DDGStatsDemocratsFeature, Set.empty[Long]).contains)),
    (
      "author_is_republican",
      candidate =>
        candidate
          .getOrElse(AuthorIdFeature, None)
          .exists(candidate.getOrElse(DDGStatsRepublicansFeature, Set.empty[Long]).contains)),
    )

Re: Twitter's Recommendation Algorithm

#13

Post by jonahbenton »

throwayyy479087 wrote: post #5

You gotta hand it to Elon - he actually did it.

LOL. My algorithm at twitter had been very simple-

See tweets from people I followed.

Don't see tweets from people I didn't follow.

Trust people I follow in their retweets to signal something interesting.

Unfollow unhelpful people.

Once that algorithm was rendered impossible, I left twitter.

Haven't missed it.

Having someone say- here's the way we are going to promote something to you- doesn't make me inclined to accept the promotion!

Re: Twitter's Recommendation Algorithm

#14

Post by mquander »

minimaxir wrote: post #9
throwayyy479087 wrote: post #5

You gotta hand it to Elon - he actually did it.

If you look at the GitHub repo, most of it is READMEs describing systems, not the models or code subleties which actually give explanations into how certain weird behaviors on Twitter happen. (e.g. the preference of certain users in the For You tab. EDIT: bad example, since there appears to be a flag for that in the code, although it does not specify which users are on the list)

The links in the README just go to other documents, but the repo seems to have most of the code for the components the documents are describing.

Re: Twitter's Recommendation Algorithm

#17

Post by sithlord »

crop_rotation wrote: post #8

Wouldn't any such system depend on 10 other internal systems, 20 databases directly or indirectly, each affecting the behaviour of the recommendation engine. That makes me doubtful studying such a recommendation engine is any better than a purely academic exercise.

thats why its "the algorithm" not the source of data/truth

Re: Twitter's Recommendation Algorithm

#18

Post by softfalcon »

crop_rotation wrote: post #8

Wouldn't any such system depend on 10 other internal systems, 20 databases directly or indirectly, each affecting the behaviour of the recommendation engine. That makes me doubtful studying such a recommendation engine is any better than a purely academic exercise.

You’re probably right, but analyzing such things could still be useful for research.

I know that open source code around commenting online directly impacted the direction my current team went building our community tooling.

I’ll take even a glimpse into the machinations of any social media giant. It’s better than nothing!

Re: Twitter's Recommendation Algorithm

#19

Post by minimaxir »

mquander wrote: post #14
minimaxir wrote: post #9

Earlier quoted context omitted.

If you look at the GitHub repo, most of it is READMEs describing systems, not the models or code subleties which actually give explanations into how certain weird behaviors on Twitter happen. (e.g. the preference of certain users in the For You tab. EDIT: bad example, since there appears to be a flag for that in the code, although it does not specify which users are on the list)

The links in the README just go to other documents, but the repo seems to have most of the code for the components the documents are describing.

It seems to vary by service; some are more detailed than others.
Post reply on HN