April 22 - 26, 2024
Seattle, Washington
May 7 - 9, 2024 (Virtual)

Event Supporters

2024 MRS Spring Meeting
MT01.05.05

End-To-End Differentiability and Tensor Processing Unit Computing (TPU) to Accelerate Materials’ Inverse Design

When and Where

Apr 24, 2024
10:30am - 11:00am
Room 320, Level 3, Summit

Presenter(s)

Co-Author(s)

Mathieu Bauchy1

University of California, Los Angeles1

Abstract

Mathieu Bauchy1

University of California, Los Angeles1
Numerical simulations have revolutionized material design. However, although simulations excel at mapping an input material to its output property, their direct application to inverse design has traditionally been limited by their high computing cost and lack of differentiability. Here, taking the example of the inverse design of a porous matrix featuring a target sorption isotherm, we introduce a computational inverse design framework that addresses these challenges. First, we adopt end-to-end differentiability to build a differentiable forward numerical simulation. Thanks to its differentiability, the forward simulation is used to directly train a backward deep generative model, which outputs an optimal porous matrix based on an arbitrary input sorption isotherm curve. Second, this inverse design pipeline leverages the power of tensor processing units (TPU)—an emerging family of dedicated chips, which, although they are specialized in deep learning, are flexible enough for intensive scientific simulations. This approach holds promise to accelerate the inverse design of novel materials with tailored properties.

Symposium Organizers

Raymundo Arroyave, Texas A&M Univ
Elif Ertekin, University of Illinois at Urbana-Champaign
Rodrigo Freitas, Massachusetts Institute of Technology
Aditi Krishnapriyan, UC Berkeley

Session Chairs

Aditi Krishnapriyan
Wennie Wang

In this Session