Articles about "Recommendation Systems"

Meta has established the MRS Research team to overhaul its advertising algorithms across platforms such as Facebook, Instagram, and Threads. Led by VP Yang Song and staffed by experts from leading AI firms, the group is focused on advancing recommendation systems and improving the efficiency of ad targeting. The initiative comes amid Meta's increased investment in AI-driven ad optimization and a forecasted rise in advertising revenue.

Meta has launched a new feature for Threads enabling users to influence their algorithmic feeds in real time by posting requests starting with 'dear algo.' The update follows months of user criticism and a viral meme, aiming to provide greater control over content recommendations. The feature is currently available in several English-speaking countries, with broader rollout expected.

X's recent release of its recommendation algorithm code has drawn criticism from researchers, who describe it as heavily redacted and of limited transparency. Despite claims of openness, experts highlight the absence of key decision-making details and the growing opaqueness resulting from neural network-based ranking. The lack of access to training data and model weightings leaves substantial gaps for those seeking to evaluate or reproduce X’s system.