Ravi Vakil, Stanford University

Title: How AI might accelerate mathematics discovery: a realist's guide

Abstract: There has been a great deal of exaggeration and sloppy thinking in the public discussion of the impact on AI on mathematics.  The most interesting and subtle question is to understand the actual current frontier --- what sorts of  problems are on the cusp of becoming tractable, and where should we focus our energies next?  Now is the time for rapid experimentation, to better understand the new landscape.  I will discuss two such types of experiments --- of seeing how  models can be used systematically by mathematicians to accelerate mathematical discovery in new fields, and how and  why to formalize mathematics in much more complicated domains than before.