cs.IR · 2026-09-21 · No. 120

Information Retrieval, 2026-09-21.

3 new papers in cs.IR. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

3 entries
  1. 01

    Predictable Failure in Multi-Hop Retrieval: Score-Distributional Confidence Scoring and Abstention

    Andre Bacellar

    cs.IR · cs.CL · cs.LG

    Multi-hop retrieval failures are not uniformly distributed across queries: they cluster in structurally predictable subpopulations. We prove two results formalizing this structure. First (CWAR Reducibility): confident-failure reduction is achievable if and only if retrieval features carry mutual information about success, a condition satisfied by LLM-judge pipelines but substantially weaker in dense-only settings, explaining the AUC-AC gap...

    arxiv.org/abs/2609.22056 · PDF

  2. 02

    Hybrid GPU-CPU Retrieval for Personalized Search at Ultra-Large Scale

    Hao Fu, Jichao Sun, Baiting Zhu, Qiaoling Liu, Yan Shi, Cheng Lu, Liu Liu, Yubo Wang, Xin Yao, Xiangyu Niu, Xu Dong,...

    cs.IR · cs.DC · cs.LG · cs.PF

    Embedding-based retrieval on user-generated content at the trillion-document scale exposes a sharp conflict between two production demands: deep, expressive personalization for queries with rich user intent, and broad coverage of a massive inventory under fixed latency and resource budgets. We characterize this as the personalization-scale paradox: hosting the full serving inventory in GPU memory is too resource intensive, while CPU compute...

    arxiv.org/abs/2609.21281 · PDF

  3. 03

    Verify, Don't Trust: Agentic Model Development for Video Discovery Retrieval at Scale

    Hao Fu, Baiting Zhu, Minglei Chen, Yinjie Huang, Shuai Ding

    cs.IR · cs.AI · cs.LG

    Large language model (LLM) agents can propose, implement, and evaluate model changes. Autoresearch loops demonstrate this capability through minutes-scale iterations on a self-contained program. Online autoresearch instead spans asynchronous systems, hours-long variants, and weeks-long campaigns that can influence a product. A completed run can still support an invalid conclusion when a code change is a no-op, data windows leak, evaluator...

    arxiv.org/abs/2609.21257 · PDF

This edition is part of The Daily Abstract — cs.IR archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.