AI for the Working Mathematician

Speaker: Jordan Ellenberg, University of Wisconsin-Madison

Title: Large Hypercubes

Abstract: I’ll talk about this paper: Bruhat intervals that are large hypercubes <arxiv.org/abs/2601.01235> in which we use a tool from Google, AlphaEvolve, to discover unexpected combinatorial structures inside the symmetric group (to be precise, a large hypercube in the Bruhat order.) I’ll talk about why this problem was a good fit for this particular machine learning protocol, which is a bit different from the most popular prevailing approaches. And I’ll also just talk about this large hypercube, which is interesting in its own right!