Discovery in Mathematics with Automated Conjecturing

Explore the cutting-edge field of automated conjecturing in mathematics through this seminar talk by Randy Davila from RelationalAI and Rice University. Delve into how artificial intelligence applies heuristic-driven methods to mathematical discovery, with systems like Fajtlowicz's Graffiti, DeLaViña's Graffiti.pc, and TxGraffiti collectively contributing to over 130 mathematical journal publications since the late 1980s. Learn about the evolution of automated conjecturing with a focus on TxGraffiti, which employs linear optimization methods and various heuristics to generate meaningful mathematical conjectures. Discover GraphMind, an innovative dueling framework where an Optimist proposes conjectures while a Pessimist seeks counterexamples, creating a feedback loop that enhances automated reasoning. The talk also introduces GraffitiAI, a Python package extending automated conjecturing across different mathematical domains. Presented by Dr. Randy Davila, a lecturer in Computational Applied Mathematics at Rice University and Library Engineer at RelationalAI who specializes in relational knowledge graph systems and holds a PhD in Mathematics from the University of Johannesburg with expertise in graph theory, combinatorial optimization, and AI applications in mathematics.