Title: Learning 4-Dimensional Knot Invariants from the Jones Polynomial

Abstract: Machine learning techniques have proven to be effective in knot
theory, identifying subtle relationships between topological invariants and
finding optimal solutions to problems with large intractable search spaces. In
this talk I will outline the current state of affairs of machine learning in
knot theory, and discuss problems of learning certain 4-dimensional topological
invariants (like the slice genus, Khovanov homology, and the Rasmussen
s-invariant) from the Jones polynomial. I will discuss what these results
suggest about the underlying structure of these 4D invariants, and identify new
lines of approach for studying knot theory via machine learning.

Talk – Slides

Talk – Video