TRACTABLE LARGE SCALE CALIBRATION WITH RL

Fadel Thior, Rose Bandolo, Sekou Remy

Keywords: Inverse Problem, Reinforcement Learning

TL;DR: RL learns differently from Bayesian Optimisation, and might provide better results for inverse problems

Abstract: In this work we show that Reinforcement Learning (RL) is an effective algorithm for calibration problems at a scale which traditionally applied Bayesian approaches struggle. This work uses synthetic data, so has access to ground truth parameters and it can be seen that RL learns different, arguably better information for different parts of the learning process. These exciting results set the foundation for deeper consideration of RL in this space.