VLDB talk: Deep Research agents need co-designed retrievers, DR Tulu and AgentIR show
CShorten30 · x · 2026-09-05
At VLDB, Akari Asai presented work on Deep Research agents, covering DR Tulu and AgentIR.
- DR Tulu: trains deep research agents with rubrics to choose between BM25 and vector search retrievers, with strong results.
- AgentIR: trains a dedicated embedding model for the agent's intermediate reasoning and search queries, a notably elegant insight.
The author suggests combining AgentIR with BM25/ColBERT-style feedback as an open research direction, bridging rationale-guided search and retrieval results.
Related event: VLDB Talk: Deep Research Agents Must Co-Train with Retrievers(3 posts)→
More from coding & agent
- Claude Code criticized: no full conversation history or search in VSCode — Birchlabs · 2026-09-05
- Tencent's 770B Hy4 preview builds a playable 3D game from one prompt in agent test — HeyToha · 2026-09-05
- Raspberry AI turns plain-English requests into fashion renders and tech packs on LangGraph — LangChain · 2026-09-05
- Perplexity CEO Declares the End of Markdown Files: Bitter Lesson Wins Again — AravSrinivas · 2026-09-05
- Codex rust-v0.153.3 Hotfix Adds GPT-6-Astra to Amazon Bedrock — github-actions[bot] · 2026-09-05
- Debunked: OpenAI Agents Didn't "Turn Evil" — They Made 15,000+ Edits to a German Wiki to Beat Task Timers — irinarish · 2026-09-05