Mean-Semivariance Policy Optimization via Risk-Averse Reinforcement Learning (Extended Abstract)
Mean-Semivariance Policy Optimization via Risk-Averse Reinforcement Learning (Extended Abstract)
Xiaoteng Ma, Shuai Ma, Li Xia, Qianchuan Zhao
Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence
Journal Track. Pages 6925-6930.
https://doi.org/10.24963/ijcai.2023/784
Keeping risk under control is often more crucial than maximizing expected rewards in real-world decision-making situations, such as finance, robotics, autonomous driving, etc. The most natural choice of risk measures is variance, while it penalizes the upside volatility as much as the downside part. Instead, the (downside) semivariance, which captures negative deviation of a random variable under its mean, is more suitable for risk-averse proposes. This paper aims at optimizing the mean-semivariance (MSV) criterion in reinforcement learning w.r.t. steady reward distribution. Since semivariance is time-inconsistent and does not satisfy the standard Bellman equation, the traditional dynamic programming methods are inapplicable to MSV problems directly. To tackle this challenge, we resort to Perturbation Analysis (PA) theory and establish the performance difference formula for MSV. We reveal that the MSV problem can be solved by iteratively solving a sequence of RL problems with a policy-dependent reward function. Further, we propose two on-policy algorithms based on the policy gradient theory and the trust region method. Finally, we conduct diverse experiments from simple bandit problems to continuous control tasks in MuJoCo, which demonstrate the effectiveness of our proposed methods.
Keywords:
Machine Learning: ML: Reinforcement learning
Machine Learning: ML: Deep reinforcement learning
Planning and Scheduling: PS: Markov decisions processes
Uncertainty in AI: UAI: Sequential decision making