mlsanityData to 7 Oct 2026

Learning Interpretable, High-Performing Policies for Autonomous Driving

arXiv 2202.02352 ↗First posted 4 Feb 2022cs.LG cs.ROcited by 11

On the map: End-to-end autonomous driving → Reinforcement learning for driving

Abstract (arXiv)

Gradient-based approaches in reinforcement learning (RL) have achieved tremendous success in learning policies for autonomous vehicles. While the performance of these approaches warrants real-world adoption, these policies lack interpretability, limiting deployability in the safety-critical and legally-regulated domain of autonomous driving (AD). AD requires interpretable and verifiable control policies that maintain high performance. We propose Interpretable Continuous Control Trees (ICCTs), a tree-based model that can be optimized via modern, gradient-based, RL approaches to produce high-performing, interpretable policies. The key to our approach is a procedure for allowing direct optimization in a sparse decision-tree-like representation. We validate ICCTs against baselines across six domains, showing that ICCTs are capable of learning interpretable policy representations that parity or outperform baselines by up to 33% in AD scenarios while achieving a 300x-600x reduction in the number of policy parameters against deep learning baselines. Furthermore, we demonstrate the interpretability and utility of our ICCTs through a 14-car physical robot demonstration.

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Placed in Reinforcement learning for driving, one of 9 directions in End-to-end autonomous driving. Builds on 32 of its 68 references that are in the map. First posted on arXiv 4 Feb 2022.

Builds on

32 of 68 references.

References · 32

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,635
  3. 2015Continuous control with deep reinforcement learningcited by 3,279
  4. 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,708
  5. 2018Addressing Function Approximation Error in Actor-Critic Methodscited by 2,302
  6. 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,031
  7. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,400
  8. 2020Mastering Atari with Discrete World Modelscited by 990
  9. 2017Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networkscited by 739
  10. 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 616
  11. 2018Data-Efficient Hierarchical Reinforcement Learningcited by 436
  12. 2019A Survey of Autonomous Driving: Common Practices and Emerging Technologiescited by 373
20 more
  1. 2021Interpretable Machine Learning: Fundamental Principles and 10 Grand Challengescited by 364
  2. 2017Distilling a Neural Network Into a Soft Decision Treecited by 230
  3. 2019Explainable Machine Learning in Deploymentcited by 210
  4. 2018Verifiable Reinforcement Learning via Policy Extractioncited by 176
  5. 2015Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction modelcited by 140
  6. 2017Interpretable Learning for Self-Driving Cars by Visualizing Causal Attentioncited by 106
  7. 2018Grounding Visual Explanationscited by 92
  8. 2017Beyond Sparsity: Tree Regularization of Deep Models for Interpretabilitycited by 68
  9. 2018Adaptive Neural Treescited by 64
  10. 2017Learning Certifiably Optimal Rule Lists for Categorical Datacited by 50
  11. 2018Generating Counterfactual Explanations with Natural Languagecited by 37
  12. 2019Robustness Verification of Tree-based Modelscited by 33
  13. 2019Interpretable and Personalized Apprenticeship Scheduling: Learning Interpretable Scheduling Policies from Heterogeneous User Demonstrationscited by 14
  14. 2019Regional Tree Regularization for Interpretability in Black Box Modelscited by 14
  15. 2019Mo"ET: Mixture of Expert Trees and its Application to Verifiable Reinforcement Learningcited by 12
  16. 2021Scalable Multiagent Driving Policies For Reducing Traffic Congestioncited by 11
  17. 2017An Optimization Approach to Learning Falling Rule Listscited by 10
  18. 2019Neural-encoding Human Experts' Domain Knowledge to Warm Start Reinforcement Learningcited by 7
  19. 2020Interpretable Control by Reinforcement Learningcited by 5
  20. 2021PyTorch, Explain! A Python library for Logic Explained Networkscited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 6; 2025: 1; 2026: 4.

11 papers.

  1. 2024Safety Implications of Explainable Artificial Intelligence in End-to-End Autonomous Drivingcited by 7
  2. 2024Advancing Explainable Autonomous Vehicle Systems: A Comprehensive Review and Research Roadmapcited by 6
  3. 2024Exploring Latent Pathways: Enhancing the Interpretability of Autonomous Driving with a Variational Autoencodercited by 6
  4. 2024Optimizing Interpretable Decision Tree Policies for Reinforcement Learningcited by 4
  5. 2024KnowPC: Knowledge-Driven Programmatic Reinforcement Learning for Zero-shot Coordinationcited by 2
  6. 2024In Search of Trees: Decision-Tree Policy Synthesis for Black-Box Systems via Searchcited by 1
  7. 2025Towards Automated Semantic Interpretability in Reinforcement Learning via Vision-Language Modelscited by 1
  8. 2026Event-Grounded Sparse Autoencoders for Vision-Language-Action Policiescited by 1
  9. 2026ORCAID: Oblique Rule-Based Continuous-Action Interpretation for Deep RL Policiescited by 0
  10. 2026ConceptTree: Bringing Semantic Transparency to Black-Box Decision Making for Robotic Manipulationcited by 0
  11. 2026Temporally Interpretable Differentiable Decision Treescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Interpretable Decision-Making for End-to-End Autonomous Driving
  2. 2021A Versatile and Efficient Reinforcement Learning Framework for Autonomous Driving
  3. 2022How to Learn from Risk: Explicit Risk-Utility Reinforcement Learning for Efficient and Safe Driving Strategies
  4. 2024DRIVE: Dependable Robust Interpretable Visionary Ensemble Framework in Autonomous Driving
  5. 2021Towards Learning Generalizable Driving Policies from Restricted Latent Representations
  6. 2025TreeIRL: Safe Urban Driving with Tree Search and Inverse Reinforcement Learning
  7. 2024Distilling Reinforcement Learning Policies for Interpretable Robot Locomotion: Gradient Boosting Machines and Symbolic Regression
  8. 2022Driver Dojo: A Benchmark for Generalizable Reinforcement Learning for Autonomous Driving