Large-scale Discovery of Scientific Hypotheses; Computation Over Expert Opinions

University Of Chicago, Chicago IL

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Our hypothesis is that we can drastically accelerate scientific discovery by collecting and making visible, comparable, and computable hypotheses, hunches, and theory-specific evidence beyond the dominant paradigm(s) of the field. Accordingly, we propose a multidisciplinary approach to harvest, formalize and compute over hypotheses and findings produced within any scientific community. Our approach unfolds in a four-stage cycle: 1) Search for scientific hunches and findings from diverse sources across the scientific community;2) Synthesize new hypotheses given prior findings and ideas and Formalize them into a composite reasoning model;3) Compute the probability of each hypothesis, given published statements and unpublished observations;4) Disseminate findings, hypotheses, and their certainties back to the scientific community in compelling ways that stimulate scientific engagement and advance.