Twitter reveals some of its source code, including its recommendation algorithm

EQUAL many times promised by Twitter CEO Elon Musk, Twitter has opened part of its source code for public testing, including the algorithm it uses to recommend tweets in a user’s timeline.

On GitHub, Twitter published two archives contains code for many of the parts that make the social network bookmarked, including the mechanism that Twitter uses to control the tweets users see on the For You timeline. In a blog post, Twitter described the move as “the first step to[ing] more transparent” at the same time”[preventing] risk” to Twitter itself and the people on the platform.

On a Twitter Spaces session today, Musk clarified:

“Our first release of the so-called algorithm will be pretty awkward and people will find a lot of bugs, but we’ll fix them very quickly,” Musk said. “Even if you don’t agree with something, at least you know why it’s there and that you’re not being secretly manipulated… The same thing, here, that we want is great example of Linux as an open source operating system… In theory, one could discover many exploits for Linux. In fact, what happens is the community identifies and fixes those vulnerabilities.”

Regarding the second point in the hedging blog post, the open source releases do not include code that provides Twitter’s ad recommendations or data used to train Twitter’s recommendation algorithm. Furthermore, they include some tutorials on how to test or actually use the code — reinforcing the idea that releases are entirely developer-focused.

“[We excluded] any code that could compromise the safety and privacy of our users or our ability to protect our platform from bad actors, including undermining our efforts to combat exploitation. Child sexual exploitation and manipulation,” Twitter wrote. It’s a bit of a mixed message just coming next week Twitter has laid off many of its safe and trusted employees. However, the company still insists that it “[took] steps to ensure that the safety and privacy of users are protected.”


A diagram showing how Twitter’s referral channel works.

Twitter says it is working on tools to manage code suggestions from the community and synchronize changes with its internal repository. Presumably, those will be made available at a future date – there’s no sign of them at the moment.

“We’ll be looking for suggestions, not just about the bug, but about how the algorithm works,” Musk said during the Spaces session. “It will be a development process. I don’t expect it to be a constant upward movement… but we are very open to what will improve the user experience.”

At first glance, the algorithm is quite complex — but not necessarily surprising in any way from a technical standpoint. It includes a variety of models, including one to detect “unsafe for work” or abusive content, determine how likely Twitter users are to interact with other users, and calculate a user’s “reputation.” Twitter users. (It’s unclear exactly what “reputation” refers to; the high-level documentation isn’t clear on that.) Several neural networks are responsible for ranking tweets and recommending accounts to follow, in when a filter element hides tweets to — pardon the jargon — “support legal compliance, improve product quality, increase user trust, protect revenue through the use of visible product handling, hard filtering, and detailed rating reduction.”

In a technique blog postTwitter revealed more about the recommendation pipeline, which it claims runs about five billion times per day:

“We strive to extract the best 1,500 tweets from a pool of hundreds of millions… Today, the For You timeline consists of 50% [tweets from people you don’t follow] and 50% [tweets from people you follow] average, although this can vary from user to user,” Twitter wrote. “Ranking” [tweets] achieved with a neural network of ~48 million parameters continuously trained on tweet interactions to optimize for positive engagement (e.g. likes, retweets, and replies).”

Of course, Twitter users don’t see all 1,500 tweets. They are filtered by content restrictions as well as other criteria and factors considered by the models, such as

Gizmodo note that one thing that doesn’t seem to be made public is the list of VIPs that Twitter pushes to its users. This week, Platformer report that Twitter has a rotating list of notable users, including YouTuber Mr. Beast and Daily Wire founder Ben Shapiro, which Twitter uses to track changes to its recommendation algorithm by increasing the visibility of these seemingly-at-will “power users.”

The source code release follows a number of controversies regarding tweaks to Twitter’s recommendation algorithm in recent months. Based on communication, in February, Musk called on Twitter’s engineers to reconfigure the algorithm so that his tweets were seen by more people. (Twitter later rolled back this change – at least somewhat.) In November, Twitter started display users get more tweets from people they don’t follow — a move that boosts the platform try before Musk’s acquisition but then reversed following backlash from users.

Source by [author_name]


News7h: Update the world's latest breaking news online of the day, breaking news, politics, society today, international mainstream news .Updated news 24/7: Entertainment, the World everyday world. Hot news, images, video clips that are updated quickly and reliably

Related Articles

Back to top button