Invited Talk - How can Machine Learning Help Mathematicians?

Can AI help mathematicians solve open problems in mathematics? In particular, problems that are currently beyond the reach of even the best mathematicians? This talk explores this question from two angles: mathematical discovery and automated theorem proving. We will study how different types of AI models can be trained to provide “AI-enhanced” intuition in mathematical questions, enabling humans to bypass the mathematical difficulty of these problems. In addition, we will discuss the use of online RL tree search designed for automated theorem proving in formal language.