Hypothesis Hunting with Evolving Networks of Autonomous Scientific Agents
This addresses the problem of hypothesis hunting in scientific datasets for researchers, offering a scalable approach to exploratory discovery, though it is incremental as it builds on existing agent-based and social network concepts.
The paper tackles the challenge of exploratory discovery in large-scale scientific datasets by introducing a framework called AScience, implemented as ASCollab, which uses LLM-based research agents that self-organize into evolving networks to produce and peer-review findings, resulting in the accumulation of expert-rated results including rediscoveries, extensions, and new proposals, such as in cancer cohorts.
Large-scale scientific datasets -- spanning health biobanks, cell atlases, Earth reanalyses, and more -- create opportunities for exploratory discovery unconstrained by specific research questions. We term this process hypothesis hunting: the cumulative search for insight through sustained exploration across vast and complex hypothesis spaces. To support it, we introduce AScience, a framework modeling discovery as the interaction of agents, networks, and evaluation norms, and implement it as ASCollab, a distributed system of LLM-based research agents with heterogeneous behaviors. These agents self-organize into evolving networks, continually producing and peer-reviewing findings under shared standards of evaluation. Experiments show that such social dynamics enable the accumulation of expert-rated results along the diversity-quality-novelty frontier, including rediscoveries of established biomarkers, extensions of known pathways, and proposals of new therapeutic targets. While wet-lab validation remains indispensable, our experiments on cancer cohorts demonstrate that socially structured, agentic networks can sustain exploratory hypothesis hunting at scale.