prescience

Summary

Can AI systems trained on the scientific record up to a fixed point in time forecast the scientific advances that follow? Such a capability could help researchers identify collaborators and impactful research directions, and anticipate which problems and methods will become central next. We introduce PreScience, a scientific forecasting benchmark that decomposes the research process into four interdependent generative tasks: collaborator prediction, prior work selection, contribution generation, and impact prediction.

Keywords

hf-dataset language-modeling · -qa parquet text datasets pandas polars mlcroissant scientific-papers arxiv citation-prediction author-prediction collaboration-prediction research-forecasting has-paper

Topics

Language Modeling, QA

Research notes

• downloads=169; likes=21