Top Information Retrieval Papers: Meta, YouTube, Tencent Explore RAG and Generative Recommendation
_reachsumit · x · 2026-08-03
The latest issue of Top Information Retrieval Papers of the Week (Vol. 167) summarizes recent research in RAG, generative recommendation, and search.
Key Papers & Directions:
- Open Retrieval Models: LightOn presents fully open dense (DenseOn) and late-interaction models trained on 665M English contrastive pretraining pairs, pushing back against closed data.
- Limits of Generative Retrieval: Meta publishes a paper arguing against generation for retrieval; Wang et al. explore what semantic IDs preserve and lose in generative recommendation.
- Agentic RAG Comparisons: Wang et al. provide a controlled comparison of lexical, dense, graph, and agentic RAG; Tencent proposes a RipGRep variant for efficient agentic search.
- Industry Recommender Systems: YouTube, Yandex, and Kuaishou share production-level insights on I/O efficiency, item embeddings, and multi-objective generative retrieval.
- Recommenders in the Agent Era: Aixin Sun publishes a position paper on recommender systems in the era of autonomous agents.
More from Research
- Simplex Diffusion Models paper talk by EPFL/DeepMind's Justin Deschenaux on Oct 19 — ArnaudDoucet1 · 2026-10-16
- Meituan LongCat team's Data World synthesizes data environments to train data agents — meituan-longcat · 2026-10-16
- FDE-Gym benchmark shows huge gap on real AI delivery tasks: top model scores 0.563 — DeepLethe · 2026-10-16
- ForcingBooster halves computation of autoregressive video diffusion with patch-wise denoising — VAST-AI · 2026-10-16
- Baidu open-sources FlashMask: up to 1.55x faster than FlashAttention-4 for long-context training — BaiduResearch · 2026-10-16
- SemEval-2027 Task 9 opens: multimodal news framing analysis across 4 languages — preslav_nakov · 2026-10-16