SciMON: Scientific Inspiration Machines Optimized for Novelty

arXiv:2305.14259v5 Announce Type: replace-cross
Abstract: We explore and enhance the ability of neural language models to generate novel scientific directions grounded in literature. Work on literature-based hypothesis generation has traditionally focused on binary link prediction -- severely limiting the expressivity of hypotheses. This line of work also does not focus on optimizing novelty. We take a dramatic departure with a novel setting in which models use as input background contexts (e.g., problems, experimental settings, goals), and output natural language ideas …

abstract arxiv binary cs.ai cs.cl cs.lg explore focus generate hypothesis inspiration language language models line link prediction literature machines novel prediction type work