Automating the Process of Energy-Technology Invention

David E Tew, Mihaela Quirk, Lipi Acharya

Keywords: Energy-Technology Invention, Productivity, ML-Enhanced Design Tools, DIFFERENTIATE

TL;DR: A framework and program to support the development of productivity-enhancing ML-based energy technology design tools is described.

Abstract:

Artificial Intelligence (AI) is facilitating transformative productivity enhancements by enabling ever-increasing levels of
automation. The Advanced Research Projects Agency-Energy (ARPA-E) of the U.S. Department of Energy seeks to leverage this capability to automate elements of the energy-technology development process in order to accelerate the realization of technological solutions to our climate crisis. This paper introduces three critical AI-enabled design capabilities--expert optimizers, low-cost high-fidelity evaluators, and inverse design tools--that are being developed for a wide range of energy applications and provides examples of two tools being developed to accelerate the discovery of new materials.