Get Started
Home
Topics
Search
Library
Research questionHow can open-ended discovery agents choose what to investigate while keeping claims calibrated to accumulated evidence?Open-ended discovery requires agents to pursue promising phenomena without a predefined question or analysis plan. As evidence accumulates over successive investigations, deciding what to explore next and how strongly to state a claim becomes difficult because emerging interpretations can exceed what the analyses support.
AI
AI Agents
Evaluation & Benchmarks
Machine Learning
Reasoning
Research Paper
Latest papersRecent research connected to this question, newest first.StatefulDiscovery: Evidence-Calibrated Claim Formation in Open-Ended Scientific DiscoveryThe source studies a discovery framework that externalizes investigation state to coordinate frontier selection, evidence acquisition, and claim adjudication. Evidence is reported from 40 real-data discovery tasks, with ablations examining structured hypotheses, local adjudication, frontier control, and persistent states.research paper · Sep 3, 2026
Related questions
How do we evaluate whether scientific agents make justified discoveries from data rather than reproduce known analyses?How can multi-agent LLM systems sustain autonomous exploration and knowledge accumulation in open-ended discovery?How can research agents refine multi-constraint answers while keeping evidence verified over long horizons?How can long-video agents choose evidence-acquisition strategies for focused, broad-coverage, or contrastive questions?