Robust Control of Scientific Simulations with Deep Reinforcement Learning

Daniel Faissol | 21-SI-001

Executive Summary

We propose advancing state-of-the-art deep reinforcement learning to increase its applicability to real-world problems by improving performance in unseen environments, adding interpretability and  diverse candidate solutions. If successful, our findings will advance breakthroughs in large-scale scientific simulations, precision medicine, and rapid response to novel pathogens, among other discoveries that rely on high-performance computer simulations.

Publications, Presentations, and Patents

Silva FL, A. Goncalves, S. Nguyen, D. Vashchenko, R. Glatt, T. Desautels, M. Landajuela, BK Petersen, and D. Faissol. "Leveraging Language Models to Efficiently Learn Symbolic Optimization Solutions." Proceedings of the Adaptive and Learning Agents Workshop ALA. Auckland, NZ. May 2022.