Next seminar:

Fabian Ruehle (Northeastern), Tuesday, Nov 29, 1:00 pm, 509 Lake Hall

Knot Theory and Machine Learning

First talk: 509 Lake, 1:00 pm

Tea: 1:50 pm

Second talk: 509 Lake, 2:10 pm

Second tea and discussion: 3:00 pm

Abstract:

I will discuss how machine learning (ML) can be applied to problems in knot theory. In particular, I will discuss an ML technique called reinforcement learning (RL) that produce verifiably correct (rather than stochastic) results. I will start with an introduction to Knot Theory, followed by an introduction to RL. I will then illustrate the application of RL to knot theory in two instances: Proving that any given knot is the trivial knot (the so-called unknotting problem), and proving that any given knot bounds a smooth disk in the 4-ball (the so-called sliceness problem).

Seminar meetings, Fall, 2022:

• Organizational meeting, Wednesday, September 7, 12 noon, 521 Lake Hall

• David Rosen (Northeastern), Tuesday, September 13, 1:15 pm, 509 Lake

Trustworthy inference methods for machine perception.

• Eva Miranda (UPC Barcelona), Tuesday, September 20, 1:15 pm, 509 Lake Hall

Where will the 29,000 rubber ducks wash up? From Turing Machines to Geometry and the Navier Stokes Conjecture.

• Mboyo Esole (Northeastern), Tuesday, October 4, 1:15 pm, 509 Lake

A brief introduction to Elliptic Fibrations and Collisions of Singularities

• Calina Copos (Northeastern), Tuesday, October 25, 1:15 pm, 509 Lake

A very brief introduction to fluid dynamics and mechanics

• Dani Alvarez-Gavela (MIT), Tuesday, November 8, 3:00 p.m. (note time!!), 509 Lake

The flexibility of caustics and its applications

• Fabian Ruehle (Northeastern Math + Physics), Tuesday, Nov 29, 1:00 pm (note time), 509 Lake, TBA