2026-10-19 · America/Los_Angeles · 论文 · #19

DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses

Scientific discovery requires developing worthwhile hypotheses into practical investigations, yet learning from this evolving research process remains challenging. We introduce an operational definition of valuable and testable research hypotheses: evidence-grounded claims about consequential uncertainties, coupled with feasible investigations, explicit decision criteria, and alternative explanations. Our DiscoveryRSI framework develops hypothesis trees through multi-agent evolutionary search, compiles selected investigations into executable Harbor tasks, and recursively learns through three…

热度 49.9 / 100；排名与评分保留该期记录。