Explainable Deep Reinforcement Learning: State of the Art and Challenges

Abstract

Interpretability, explainability and transparency are key issues to introducing Artificial Intelligence methods in many critical domains: This is important due to ethical concerns and trust issues strongly connected to reliability, robustness, auditability and fairness, and has important consequences towards keeping the human in the loop in high levels of automation, especially in critical cases for decision making, where both (human and the machine) play important roles. While the research community has given much attention to explainability of closed (or black) prediction boxes, there are tremendous needs for explainability of closed-box methods that support agents to act autonomously in the real world. Reinforcement learning methods, and especially their deep versions, are such closed-box methods. In this article we aim to provide a review of state of the art methods for explainable deep reinforcement learning methods, taking also into account the needs of human operators - i.e., of those that take the actual and critical decisions in solving real-world problems. We provide a formal specification of the deep reinforcement learning explainability problems, and we identify the necessary components of a general explainable reinforcement learning framework. Based on these, we provide a comprehensive review of state of the art methods, categorizing them in classes according to the paradigm they follow, the interpretable models they use, and the surface representation of explanations provided. The article concludes identifying open questions and important challenges.

References in corpus (13)

• Distilling the Knowledge in a Neural Network
• Interpretable machine learning: definitions, methods, and applications
• Not Just a Black Box: Learning Important Features Through Propagating Activation Differences
• Measuring the Quality of Explanations: The System Causability Scale (SCS). Comparing Human and Machine Explanations
• Verifiable Reinforcement Learning via Policy Extraction
• Interestingness Elements for Explainable Reinforcement Learning: Understanding Agents' Capabilities and Limitations
• Hierarchical and Interpretable Skill Acquisition in Multi-task Reinforcement Learning
• Evaluating Saliency Map Explanations for Convolutional Neural Networks: A User Study
• Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learning
• Autonomous Self-Explanation of Behavior for Interactive Reinforcement Learning Agents
• Automatic Discovery of Interpretable Planning Strategies
• What Did You Think Would Happen? Explaining Agent Behaviour Through Intended Outcomes
• Explaining by Imitating: Understanding Decisions by Interpretable Policy Learning

Cited by in corpus (7)

• Explainable Artificial Intelligence: A Survey of Needs, Techniques, Applications, and Future Direction
• Multi-Agent Reinforcement Learning: Methods, Applications, Visionary Prospects, and Challenges
• A Comprehensive Survey on Self-Interpretable Neural Networks
• How to Control Hydrodynamic Force on Fluidic Pinball via Deep Reinforcement Learning
• A survey of air combat behavior modeling using machine learning
• Compositional Concept-Based Neuron-Level Interpretability for Deep Reinforcement Learning
• Themis: An explainable AI-enabled framework for Reinforcement Learning with Human Feedback