Risk-Sensitive RL

RL: Reinforcement Learning

Overview

Risk-sensitive reinforcement learning studies how agents can make sequential decisions that account for rare failures and undesirable outcomes, not just average reward. Common approaches optimize CVaR, quantile, variance, expectile, or exponential-utility objectives, often using distributional critics, policy gradients, Q-learning, or model-based planning. Safety-constrained methods add tools such as shields, reachability checks, and robust uncertainty sets, with applications ranging from robot navigation and autonomous driving to finance and language-model alignment. The field is developing convergence and sample-complexity guarantees alongside practical methods, while a central tension remains how to estimate tail risk reliably from limited data without making policies excessively conservative.

Latest papers

CardsList

97 papers8 papers in the last four weeks, down 60% on the four weeks before. 0.1% of all new papers.

  1. Cantelli Constrained Policy Optimization

    Jan 30, 2026Rohan Tangri, Jan-Peter CalliessPolicy OptimizationConstrained RL

  2. Entropic Risk-Aware Monte Carlo Tree Search

    Jan 25, 2026Pedro P. Santos, Jacopo Silvestrin, Manuel Lopes +2Monte Carlo Tree SearchMarkov Decision Processes

  3. Risk-Aware Alignment Must Incorporate Reward Uncertainty

    Oct 31, 2024Debangshu Banerjee, Kintan Saha, Aditya GopalanPolicy OptimizationLLM Alignment

  4. TRAM: Test-Time Risk Adaptation with Mixture of Agents

    Aug 16, 2024Mohamad Chehade, Amrit Singh Bedi, Souradip Chakraborty +2Reinforcement LearningTest-Time RL

  5. ProSpec RL: Plan Ahead, then Execute

    Jul 31, 2024Liangliang Liu, Yi Guan, BoRan Wang +5Reinforcement LearningRisk-Sensitive RL

  6. On the Approximation and Convergence of Distributional Policy Gradient Algorithms for Risk-Sensitive Reinforcement Learning

    May 23, 2024Xian Yu, Minheng Xiao, Lei YingStochastic OptimizationDistributional RL