Using AI to Predict Reproducibility
Being able to reproduce the results of a scientific paper is critical, but doing so is also costly and time consuming. Even determining which papers are most likely to be able to be replicated can take significant effort from experts in the field.
Brian Uzzi and a team at Kellogg School of Management trained a neural network model to evaluate the text of journal papers to predict whether the work would be replicable. The AI model was able to predict replicability as accurately as expert prediction markets, which can take as long as a year to complete, in much less time and effort. This could give researchers and journal editors a powerful tool to aid in the review process of scientific work.
Read the full Kellogg Insight article
Learn more about Northwestern's use of AI in research exploration
Back to top