Learning to Alleviate Familiarity Bias in Video Recommendation Conference: Companion Proceedings of the ACM Web Conference 2026; April 13–17, 2026; Dubai, United Arab Emirates. Companion Proceedings of the ACM Web Conference 2026 (WWW Companion ’26), April 13–17, 2026, Dubai, United Arab Emirates CC
Zheng Ren, Yi Wu, Jianan Lu, Acar Ary, Yiqu Liu, Li Wei, Lukasz Heldt · 2026
Read Learning to Alleviate Familiarity Bias in Video Recommendation Conference: Companion Proceedings of the ACM Web Conference 2026; April 13–17, 2026; Dubai, United Arab Emirates. Companion Proceedings of the ACM Web Conference 2026 (WWW Companion ’26), April 13–17, 2026, Dubai, United Arab Emirates CC by Zheng Ren, Yi Wu, Jianan Lu, Acar Ary, Yiqu Liu, Li Wei, Lukasz Heldt (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-07987.txt.
Description
Video recommendations tend to show users what they already know, crowding out new creators and content. Zheng Ren and six co-authors present LAFB, a lightweight, model-agnostic post-ranking method that learns to alleviate familiarity bias. It models user-content familiarity from interaction features and adjusts predicted ratings with personalized debiasing factors.
Large-scale offline tests and online A/B testing in a real recommendation system show LAFB increases novel watch-time share and exposure for emerging creators and overall diversity. Overall watch time and short-term satisfaction stay stable. The method has already launched in production post-ranking. Recommendation teams gain a deployable fix for a shape imbalance that engagement tuning alone tends to worsen.
Catalog record
| CREATOR | Zheng Ren, Yi Wu, Jianan Lu, Acar Ary, Yiqu Liu, Li Wei, Lukasz Heldt |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07987