From Quantum Dynamics to Intelligent Actions: A Framework for Quantum-Accelerated Simulation in Reinforcement Learning

Artificial intelligence is transforming how complex systems like power grids and advanced robotics are controlled. However, training these systems is computationally intensive. This project solves this bottleneck by harnessing the power of quantum computing. By mapping complex systems to quantum circuits, the research drastically speeds up the training process for artificial intelligence. This work directly supports national priorities in artificial intelligence and quantum science by developing new algorithms that integrate machine learning problems with quantum hardware. Ultimately, the project benefits society by enabling the design of safe and resilient technologies. It also supports workforce development by bringing quantum and machine learning concepts into new courses and research programs, providing student research opportunities, and engaging the public through outreach events. The key objective of this project is to develop a systematic framework that accelerates reinforcement learning for systems governed by ordinary and differential-algebraic equations. The central research novelty is a new procedure, termed Moment-Closed Differential Algebra, to transform nonlinear physics-based models into a linear mapping. The approach embeds these trajectories into a time-dependent Schrodinger equation while preserving system constraints, and then reformulates policy iteration as a bilinear quantum control problem to enable highly efficient updates. By reducing the dominant costs of model-based reinforcement learning to quantum operations like Hamiltonian simulation and amplitude estimation, the methods achieve favorable dimension scaling and potential exponential speedups. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2552687 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Xiantao Li | Institution: Pennsylvania State Univ University Park, UNIVERSITY PARK, PA | Award Amount: $300,000 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2552687 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2552687.html