Robust Reinforcement Learning with Dynamic Distortion Risk Measures
This work addresses risk management in reinforcement learning for applications like finance, though it appears incremental by combining existing concepts of robustness and risk measures.
The paper tackles robust risk-aware reinforcement learning by developing a framework that accounts for environmental uncertainty and risk preferences using dynamic robust distortion risk measures, and demonstrates its performance on a portfolio allocation example.
In a reinforcement learning (RL) setting, the agent's optimal strategy heavily depends on her risk preferences and the underlying model dynamics of the training environment. These two aspects influence the agent's ability to make well-informed and time-consistent decisions when facing testing environments. In this work, we devise a framework to solve robust risk-aware RL problems where we simultaneously account for environmental uncertainty and risk with a class of dynamic robust distortion risk measures. Robustness is introduced by considering all models within a Wasserstein ball around a reference model. We estimate such dynamic robust risk measures using neural networks by making use of strictly consistent scoring functions, derive policy gradient formulae using the quantile representation of distortion risk measures, and construct an actor-critic algorithm to solve this class of robust risk-aware RL problems. We demonstrate the performance of our algorithm on a portfolio allocation example.