The 16th Atul Vyas Memorial Lecture in Mathematics (Teal Witter, CMC)

Atul Vyas was an outstanding CMC student who was majoring in Mathematics and Physics. He tragically lost his life in a train crash that occurred on September 12, 2008 in Chatsworth, California. The Mathematical Sciences Department at CMC fondly remembers Atul as someone who was equally excited by the power of mathematical abstraction and the possibilities for its applications.

In memory of Atul, the CMC Mathematical Sciences Department hosts a yearly lecture series, aimed at a general audience, on the Creative Application of Abstract Mathematical Ideas.

A brief reception will take place prior to the talk at 4:00 PM

For more details, please see the attached Flyer

Speaker: R. Teal Witter, Assistant Professor of Mathematical and Computer Science, CMC

Title: Estimating Shapley Values for Explainable AI via Richer Model Approximations

Abstract: Gradient descent is at the heart of modern machine learning: We iteratively update the weights of machine learning models to minimize a problem-specific loss. When it works well, we deploy the model in human-facing domains like healthcare, finance, or the justice system. But even though we know how models are trained, we don’t understand why they make the decisions they do. A particularly compelling approach to explaining AI predictions is the Shapley value, a game-theoretic quantity that measures how each input to the model affects its output. Mathematically, the i-th Shapley value is the average change in the i-th dimension of a particular function defined on the d-dimensional hypercube. Because the hypercube has 2^d points, exactly computing Shapley values is infeasible. In this talk, we will instead leverage algorithmic insights to develop state-of-the-art approximation methods.