🤖 AI Summary
This work addresses the critical limitations of large language models (LLMs) in scientific hypothesis generation—namely, insufficient granularity and lack of experimental feasibility. We formally define, for the first time, the task of *fine-grained scientific hypothesis discovery* in chemistry: given a broad research direction, generate hypotheses specifying concrete methodologies, controllable variables, and empirically testable predictions. To this end, we propose a hierarchical search framework that jointly optimizes prompt structure, implicit reward modeling via LLM self-evaluation, multi-model ensemble scoring, and combinatorial optimization. Evaluated on a novel, expert-annotated chemical hypothesis benchmark, our approach significantly outperforms strong baselines. Results demonstrate that hierarchical search effectively smooths the reward landscape, enhancing search stability and hypothesis quality. This work reveals the upper bounds of LLMs in structured scientific reasoning and establishes a foundation for experimentally grounded hypothesis generation.
📝 Abstract
Large language models (LLMs) have shown promise in automating scientific hypothesis generation, yet existing approaches primarily yield coarse-grained hypotheses lacking critical methodological and experimental details. We introduce and formally define the novel task of fine-grained scientific hypothesis discovery, which entails generating detailed, experimentally actionable hypotheses from coarse initial research directions. We frame this as a combinatorial optimization problem and investigate the upper limits of LLMs' capacity to solve it when maximally leveraged. Specifically, we explore four foundational questions: (1) how to best harness an LLM's internal heuristics to formulate the fine-grained hypothesis it itself would judge as the most promising among all the possible hypotheses it might generate, based on its own internal scoring-thus defining a latent reward landscape over the hypothesis space; (2) whether such LLM-judged better hypotheses exhibit stronger alignment with ground-truth hypotheses; (3) whether shaping the reward landscape using an ensemble of diverse LLMs of similar capacity yields better outcomes than defining it with repeated instances of the strongest LLM among them; and (4) whether an ensemble of identical LLMs provides a more reliable reward landscape than a single LLM. To address these questions, we propose a hierarchical search method that incrementally proposes and integrates details into the hypothesis, progressing from general concepts to specific experimental configurations. We show that this hierarchical process smooths the reward landscape and enables more effective optimization. Empirical evaluations on a new benchmark of expert-annotated fine-grained hypotheses from recent chemistry literature show that our method consistently outperforms strong baselines.