VLDB talk: DR Tulu and AgentIR argue deep research needs co-designed agents and retrievers
CShorten30 · x · 2026-09-05
Akari Asai's VLDB talk covered DR Tulu and AgentIR, arguing deep research demands co-designing agents and retrievers. Key threads: training agents end-to-end for open-ended, hard-to-verify tasks — DR Tulu uses rubrics to train agents to leverage BM25 or vector retrievers with strong results — and letting retrievers exploit an agent's reasoning rather than just its latest query, via AgentIR's embedding model trained on intermediate reasoning plus search queries. The sharer suggests combining AgentIR with BM25/ColBERT-style feedback as a promising direction.
Related event: VLDB talk: Deep Research agents need co-designed retrievers(4 posts)→
More from coding & agent
- JSON prompting with MOSS Image 5B: bbox coordinates for exact multi-room layouts — cocktailpeanut · 2026-10-21
- Rewriting an agent skill: A/B challenge runs show skills should only add what agents don't know — jmschreiber91 · 2026-10-21
- Developer builds himself a 'Rork MoTS' agent on Grok Bot, aiming for full autonomy before GTA 6 — rudrank · 2026-10-21
- OpenAI Live API tip: control voice sessions from your backend via sideband WebSocket — pbbakkum · 2026-10-21
- OpenAI's MultiLangBench: PHP scores highest (91.3%) at building a database engine from scratch — ricklamers · 2026-10-21
- UserProof lets Claude Code watch your user session recordings and boost conversion for $29/mo — doooyle · 2026-10-21