Project Grant 2617987

The National Science Foundation Division of Computing and Communication Foundations awarded The Johns Hopkins University $206,849 under the Computer and Information Science and Engineering program (CFDA 47.070) on October 1, 2025, for research on causal reasoning using information-theoretic methods. The project runs through December 31, 2027, with place of performance in Baltimore, Maryland.

The research develops algorithms that identify information-theoretically simple explanations of causal systems from data, extending the applicability of causal reasoning theory beyond worst-case scenarios to real-world systems that violate classical assumptions. The first research thrust develops methods to learn causal relations from observational data through an information-theoretic interpretation of Occam's Razor based on system entropy. The second thrust analyzes how information-theoretically simple explanations enable approximate computation of causal effects that are not identifiable in the worst case. The third thrust leverages these results to develop experimental design algorithms for efficiently learning causal structures. This work addresses a fundamental problem in artificial intelligence and has applications across engineering, computer science, and medical research.