mlsanityData to 14 Oct 2026

The Interpretability of Codebooks in Model-Based Reinforcement Learning is Limited

arXiv 2407.19532 ↗First posted 28 Jul 2024cs.AI cs.LGcited by 0

On the map: Deep reinforcement learning methods → Interpretable and logic-guided RL

Abstract (arXiv)

Interpretability of deep reinforcement learning systems could assist operators with understanding how they interact with their environment. Vector quantization methods -- also called codebook methods -- discretize a neural network's latent space that is often suggested to yield emergent interpretability. We investigate whether vector quantization in fact provides interpretability in model-based reinforcement learning. Our experiments, conducted in the reinforcement learning environment Crafter, show that the codes of vector quantization models are inconsistent, have no guarantee of uniqueness, and have a limited impact on concept disentanglement, all of which are necessary traits for interpretability. We share insights on why vector quantization may be fundamentally insufficient for model interpretability.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Interpretable and logic-guided RL, one of 15 directions in Deep reinforcement learning methods. Builds on 28 of its 38 references that are in the map. First posted on arXiv 28 Jul 2024.

Builds on

28 of 38 references.

References · 28

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  3. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,733
  4. 2017Neural Discrete Representation Learningcited by 4,712
  5. 2021Zero-Shot Text-to-Image Generationcited by 4,288
  6. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,253
  7. 2018World Modelscited by 1,513
  8. 2018Sanity Checks for Saliency Mapscited by 1,156
  9. 2020Mastering Atari with Discrete World Modelscited by 1,007
  10. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
  11. 2019Attention is not Explanationcited by 929
  12. 2019Attention is not not Explanationcited by 636
16 more
  1. 2022Transformers are Sample-Efficient World Modelscited by 328
  2. 2021Benchmarking the Spectrum of Agent Capabilitiescited by 200
  3. 2022Toward Transparent AI: A Survey on Interpreting the Inner Structures of Deep Neural Networkscited by 129
  4. 2017Unsupervised Learning of Disentangled and Interpretable Representations from Sequential Datacited by 93
  5. 2023Codebook Features: Sparse and Discrete Interpretability for Neural Networkscited by 48
  6. 2021Learning Disentangled Representations in the Imaging Domaincited by 46
  7. 2023Disentanglement via Latent Quantizationcited by 40
  8. 2020Privacy-preserving Voice Analysis via Disentangled Representationscited by 24
  9. 2021What Can I Do Here? Learning New Skills by Imagining Visual Affordancescited by 23
  10. 2022Vector Quantized Image-to-Image Translationcited by 16
  11. 2020Controllable Text Generation with Focused Variationcited by 7
  12. 2023Neural Radiance Field Codebookscited by 7
  13. 2020Evidence-Aware Inferential Text Generation with Vector Quantised Variational AutoEncodercited by 6
  14. 2020Smaller World Models for Reinforcement Learningcited by 4
  15. 2022Deep Reinforcement Learning with Vector Quantized Encodingcited by 2
  16. 2023Improving Explainability of Disentangled Representations using Multipath-Attribution Mappingscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Towards Automated Semantic Interpretability in Reinforcement Learning via Vision-Language Models
  2. 2025Balancing Interpretability and Performance in Reinforcement Learning: An Adaptive Spectral Based Linear Approach
  3. 2025From Explainability to Interpretability: Interpretable Policies in Reinforcement Learning Via Model Explanation
  4. 2025Compositional Concept-Based Neuron-Level Interpretability for Deep Reinforcement Learning
  5. 2023Harnessing Discrete Representations For Continual Reinforcement Learning
  6. 2022Adaptive Discrete Communication Bottlenecks with Dynamic Vector Quantization
  7. 2022Towards Interpretable Deep Reinforcement Learning Models via Inverse Reinforcement Learning
  8. 2026Same Predictions, Different Reasons: The Effect of Quantization on Model Explanations