Autonomous Pressure Control in MuVacAS via Deep Reinforcement Learning and Deep Learning Surrogate Models
Published in Machine Learning and the Physical Sciences Workshop, NeurIPS 2025, 2025
Recommended citation: G. Rodriguez-Llorente, G. Gallardo Romero, R. Morant Navascués, N. Khvatkin Petrovsky, A. Sabogal, and R. Gómez-Espinosa Martín. Autonomous Pressure Control in MuVacAS via Deep Reinforcement Learning and Deep Learning Surrogate Models. Machine Learning and the Physical Sciences Workshop, NeurIPS 2025. arXiv:2512.15521. https://arxiv.org/abs/2512.15521
This work presents a data-driven approach to autonomous pressure control in the MuVacAS experimental facility. A deep-learning surrogate model is trained using operational data and used as a fast simulation environment for training a deep reinforcement-learning controller.
My contribution is part of the MuVacAS research programme, connecting experimental vacuum engineering with data-driven modelling, digital twins and advanced control.