mlsanityData to 14 Oct 2026

Hierarchical Support Vector State Partitioning for Distilling Black Box Reinforcement Learning Policies

arXiv 2605.04254 ↗First posted 5 May 2026cs.LG cs.HCcited by 0

On the map: Deep reinforcement learning methods → Deep RL algorithms and representations

Abstract (arXiv)

We introduce State Vector Space Partitioning (SVSP), a novel method to mimic a black box reinforcement learning policy using a set of human-interpretable subpolicies. By partitioning a distillation dataset of state action pairs with linear support vector machine splits, SVSP constructs a compact and structured representation of the original policy. Our method improves mean return by +7.4% over previous critic driven state partitioning attempts such as Voronoi State Partitioning (VSP) and +2.8% over the original TD3 policy, while reducing the number of required subpolicies against VSP by 82.1%. Our results pave the path towards a more flexible form of distillation where both the decision boundary and surrogate models can be chosen within a margin of the original black box behavior.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods. Builds on 6 of its 8 references that are in the map. First posted on arXiv 5 May 2026.

Builds on

6 of 8 references.

References · 6

  1. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,070
  2. 2018Verifiable Reinforcement Learning via Policy Extractioncited by 177
  3. 2024Interpretable and Editable Programmatic Tree Policies for Reinforcement Learningcited by 18
  4. 2021Multiclass Optimal Classification Trees with SVM-splitscited by 6
  5. 2024Human-Readable Programs as Actors of Reinforcement Learning Agents Using Critic-Moderated Evolutioncited by 1
  6. 2025Explainable RL Policies by Distilling to Locally-Specialized Linear Policies with Voronoi State Partitioningcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Critic-Driven Voronoi-Quantization for Distilling Deep RL Policies to Explainable Models
  2. 2022Deep Reinforcement Learning with Vector Quantized Encoding
  3. 2024Symbolic State Partitioning for Reinforcement Learning
  4. 2025A Survey of State Representation Learning for Deep Reinforcement Learning
  5. 2025Refined Policy Distillation: From VLA Generalists to RL Experts
  6. 2022Behavior Prior Representation learning for Offline Reinforcement Learning
  7. 2025Distilling Reinforcement Learning into Single-Batch Datasets
  8. 2026Vector Symbolic Policy Gradient