VLDB Talk: Deep Research Agents Must Co-Train with Retrievers

In a VLDB talk, Akari Asai presented DR Tulu and AgentIR, arguing that Deep Research agents should be trained jointly with retrievers, letting agents learn to choose retrieval tools via rubric-based training.

2026-09-05 ~ 2026-09-05 · 3 related posts

2 near-duplicate retellings: CShorten30 · CShorten30