Review of Metrics to Measure the Stability, Robustness and Resilience of Reinforcement Learning
Not on the map. Why?
Abstract (arXiv)
Reinforcement learning has received significant interest in recent years, due primarily to the successes of deep reinforcement learning at solving many challenging tasks such as playing Chess, Go and online computer games. However, with the increasing focus on reinforcement learning, applications outside of gaming and simulated environments require understanding the robustness, stability, and resilience of reinforcement learning methods. To this end, we conducted a comprehensive literature review to characterize the available literature on these three behaviors as they pertain to reinforcement learning. We classify the quantitative and theoretical approaches used to indicate or measure robustness, stability, and resilience behaviors. In addition, we identified the action or event to which the quantitative approaches were attempting to be stable, robust, or resilient. Finally, we provide a decision tree useful for selecting metrics to quantify the behaviors. We believe that this is the first comprehensive review of stability, robustness and resilience specifically geared towards reinforcement learning.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 107 of its 232 references that are in the map. First posted on arXiv 22 Mar 2022.
Builds on
107 of 232 references.
References · 107
- 2020Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixelscited by 573
- 2020Learning Invariant Representations for Reinforcement Learning without Reconstructioncited by 424
- 2020Responsive Safety in Reinforcement Learning by PID Lagrangian Methodscited by 302
- 2019Improving Sample Efficiency in Model-Free Reinforcement Learning from Imagescited by 268
- 2020Robust Deep Reinforcement Learning against Adversarial Perturbations on State Observationscited by 235
- 2019Minimax Weight and Q-Function Learning for Off-Policy Evaluationcited by 150
- 2019Double Reinforcement Learning for Efficient Off-Policy Evaluation in Markov Decision Processescited by 118
- 2019Ego-Pose Estimation and Forecasting as Real-Time PD Controlcited by 87
- 2018Reinforcement Learning with Perturbed Rewardscited by 76
- 2019Robust Reinforcement Learning for Continuous Control with Model Misspecificationcited by 76
- 2020Robust Deep Reinforcement Learning through Adversarial Losscited by 73
- 2019Corruption-robust exploration in episodic reinforcement learningcited by 72
38 more
- 2017Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedbackcited by 71
- 2020Automatic Data Augmentation for Generalization in Deep Reinforcement Learningcited by 71
- 2018Safe Reinforcement Learning with Model Uncertainty Estimatescited by 70
- 2019ROBEL: Robotics Benchmarks for Learning with Low-Cost Robotscited by 70
- 2020Robust Reinforcement Learning using Adversarial Populationscited by 64
- 2020Multimodal Safety-Critical Scenarios Generation for Decision-Making Algorithms Evaluationcited by 54
- 2019Sub-policy Adaptation for Hierarchical Reinforcement Learningcited by 47
- 2019Certified Adversarial Robustness for Deep Reinforcement Learningcited by 47
- 2020Robust Reinforcement Learning via Adversarial training with Langevin Dynamicscited by 45
- 2021A Deep Value-network Based Approach for Multi-Driver Order Dispatchingcited by 45
- 2020ALLSTEPS: Curriculum-driven Learning of Stepping Stone Skillscited by 44
- 2020Overcoming Model Bias for Robust Offline Deep Reinforcement Learningcited by 42
- 2020Q-Learning in Regularized Mean-field Gamescited by 40
- 2019Policy Optimization for H2 Linear Control with H_∞ Robustness Guarantee: Implicit Regularization and Global Convergencecited by 39
- 2020Distributional Robustness and Regularization in Reinforcement Learningcited by 39
- 2020On the Robustness of Cooperative Multi-Agent Reinforcement Learningcited by 39
- 2020Actor-Critic Reinforcement Learning for Control with Stability Guaranteecited by 39
- 2020Optimizing for the Future in Non-Stationary MDPscited by 37
- 2020Towards Open Ad Hoc Teamwork Using Graph-based Policy Learningcited by 37
- 2020Robustifying Reinforcement Learning Agents via Action Space Adversarial Trainingcited by 30
- 2020Improving Robustness via Risk Averse Distributional Reinforcement Learningcited by 28
- 2020S3K: Self-Supervised Semantic Keypoints for Robotic Manipulation via Multi-View Consistencycited by 28
- 2020Mitigating Value Hallucination in Dyna Planning via Multistep Predecessor Modelscited by 27
- 2019DoorGym: A Scalable Door Opening Environment And Baseline Agentcited by 26
- 2017A Deep Policy Inference Q-Network for Multi-Agent Systemscited by 24
- 2019Long-Range Indoor Navigation with PRM-RLcited by 24
- 2020Statistically Efficient Off-Policy Policy Gradientscited by 24
- 2020Training Adversarial Agents to Exploit Weaknesses in Deep Control Policiescited by 24
- 2020A General Framework for Learning Mean-Field Gamescited by 24
- 2020Multi-Agent Deep Reinforcement Learning for HVAC Control in Commercial Buildingscited by 24
- 2019CAQL: Continuous Action Q-Learningcited by 22
- 2016Riemannian game dynamicscited by 18
- 2020Robust Reinforcement Learning with Wasserstein Constraintcited by 18
- 2020Robust Spammer Detection by Nash Reinforcement Learningcited by 17
- 2019Gossip-based Actor-Learner Architectures for Deep Reinforcement Learningcited by 16
- 2019Robust Model-free Reinforcement Learning with Multi-objective Bayesian Optimizationcited by 16
- 2020t-Soft Update of Target Network for Deep Reinforcement Learningcited by 16
- 2019Learning Variable Impedance Control for Contact Sensitive Taskscited by 15
50 most cited of 107 All 107 (JSON)
Built on by
Citing papers per year: 2025: 2.
2 papers.