A new framework called Supersession-Decay Filtering (SDF) has been developed and deployed in Google Discover to combat stale recommendations. This system addresses staleness through two primary mechanisms: supersession, where new information makes older content obsolete, and relevance decay, where content naturally loses value over time. Online experiments and a two-year production deployment showed a significant reduction in user-filed staleness reports, indicating improved user engagement and a more robust method for managing content relevance at scale. AI
IMPACT This framework offers a scalable solution to improve user engagement by reducing stale content in large-scale recommendation systems.
RANK_REASON The cluster contains a research paper detailing a new framework for recommender systems, including its deployment in a major product.
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- Connected Papers
- CORE Recommender
- DagsHub
- Google Discover
- Gotit.pub
- Hugging Face
- Influence Flower
- Litmaps
- ScienceCast
- scite Smart Citations
- Supersession-Decay Filtering
- Signed Directional Distance Function
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →