Twitter announced on Friday that it has open sourced the code behind the recommendation algorithm the platform uses to select content from users’ For You timeline.

However, the code released today does not include the parties behind the ad recommendations, or it would jeopardize Twitter’s ability to keep threat actors’ attempts to manipulate the platform in check.

“For this release, we aimed for the highest degree of transparency possible, while excluding any code that would compromise user security and privacy or the ability to protect our platform from bad actors, including undermining our efforts to fight against the sexual exploitation and manipulation of children”, the company said.

“Today’s release also does not include the code that powers our ad recommendations. We’ve also taken additional steps to ensure that users’ security and privacy are protected, including our decision not to post any training data or model weights associated with the Twitter algorithm at this time.”

Twitter has released two separate GitHub repositories containing the source code for its recommendation algorithm And some of the machine learning (ML) models feeding it.

Most of the recommendation algorithm will be made open source today. The rest will follow.

The litmus test is that independent third parties should be able to determine, with reasonable accuracy, what is likely to be shown to users.

No doubt many embarrassing problems will be… https://t.co/41U4oexIev

— Elon Musk (@elonmusk) March 31, 2023

As revealed by the company’s engineering team, tweets that end up in the For You timeline are chosen by a service called Home Mixer which uses the following pipeline:

1. Grab top Tweets from different recommendation sources in a process called search for candidates.
2. Rank each Tweet using a machine learning model.
3. Apply heuristics and filterssuch as filtering Tweets from users you’ve blocked, NSFW content, and Tweets you’ve already seen.

“For each request, we try to pull the top 1,500 Tweets from a pool of hundreds of millions via these sources,” Twitter explains.

“We find candidates among people you follow (In-Network) and among people you don’t follow (Out-of-Network).”

The end goal is for each user’s For You timeline to show 50% of relevant and recent tweets from their followers and the remaining 50% from people outside their network based on what the user would find interesting.

Twitter source code leaked online months ago

Earlier this month, Twitter removed proprietary source code and internal tools leaked on GitHub and accessible to the public for at least several months.

In a DMCA Infringement Noticethe company also asked GitHub to provide access history information for the leaked code, which could uncover who downloaded the code while it was available online.

Twitter is also trying to use a subpoena filed in the United States District Court for the Northern District of California to force GitHub to share identifying information about the FreeSpeechEnthusiasm user who first posted the files and anyone who accessed and distributed the leaked Twitter source code, which could also be used for further legal action.

Today’s announcement follows tweets from Twitter CEO Elon Musk promising to go public with Twitter’s algorithm.

The first is a survey (as of March 24, 2022) which asked users to vote on a poll to decide whether “the Twitter algorithm should be open source” and the second (as of March 17, 2023) said “Twitter will open source all code used to recommend tweets on March 31.”