π€ AI Summary
This work addresses the superficiality often observed in scientific hypothesis generation for drug discovery, which stems from the vast search space and insufficient support for usersβ reasoning processes. To overcome this limitation, the authors propose a human-AI collaborative co-abduction framework and implement it in HALO, a system that introduces co-abduction into interactive abductive reasoning for the first time. HALO integrates candidate molecule clustering, strategy identification, and multi-strategy fusion to enable deep, human-in-the-loop hypothesis generation. In experiments with ten medicinal chemists, HALO significantly enhanced both the quality of generated hypotheses and molecular diversity, effectively facilitating efficient observation, systematic strategy formulation, and coherent integration of multiple reasoning strategies.
π Abstract
Scientific discovery is essential yet inefficient, primarily because generating hypotheses within a vast search space hinders breakthroughs. While current AI systems assist in generating new hypothesis candidates, they lack interactive support for the reasoning process by which users develop these outputs into promising hypotheses, resulting in surface-level hypotheses. To address this issue, we present co-abduction, a human-AI collaborative framework for abductive reasoning in scientific hypothesis generation. To operationalize co-abduction, we build HALO, a human-AI collaborative system for molecular hypothesis generation in drug discovery, enabling improved candidate clustering, strategy identification, and multi-strategy synthesis. In expert studies involving 10 medicinal chemists, HALO significantly facilitated abductive reasoning for hypothesis generation -- efficient candidate observation, systematic strategy identification, and coherent multi-strategy composition -- and enabled participants to produce higher-quality, more diverse candidate molecules.