mlsanityData to 13 Oct 2026

Unsupervised Causal Binary Concepts Discovery with VAE for Black-box Model Explanation

arXiv 2109.04518 ↗First posted 9 Sep 2021cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

We aim to explain a black-box classifier with the form: `data X is classified as class Y because X textithas A, B and textitdoes not have C' in which A, B, and C are high-level concepts. The challenge is that we have to discover in an unsupervised manner a set of concepts, i.e., A, B and C, that is useful for the explaining the classifier. We first introduce a structural generative model that is suitable to express and discover such concepts. We then propose a learning process that simultaneously learns the data distribution and encourages certain concepts to have a large causal influence on the classifier output. Our method also allows easy integration of user's prior knowledge to induce high interpretability of concepts. Using multiple datasets, we demonstrate that our method can discover useful binary concepts for explanation.

Key facts. Cited by 8 papers in the map (data through 13 Oct 2026). Builds on 23 of its 39 references that are in the map. First posted on arXiv 9 Sep 2021.

Builds on

23 of 39 references.

References · 23

  1. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,713
  2. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,328
  3. 2013Auto-Encoding Variational Bayescited by 4,024
  4. 2017Axiomatic Attribution for Deep Networkscited by 3,368
  5. 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,316
  6. 2016The Concrete Distribution: A Continuous Relaxation of Discrete Random Variablescited by 1,142
  7. 2014Striving for Simplicity: The All Convolutional Netcited by 989
  8. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 934
  9. 2018Understanding disentangling in β-VAEcited by 334
  10. 2018Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negativescited by 262
  11. 2019On Completeness-aware Concept-Based Explanations in Deep Neural Networkscited by 239
  12. 2017CausalGAN: Learning Causal Implicit Generative Models with Adversarial Trainingcited by 152
11 more
  1. 2019Explaining Classifiers with Causal Concept Effect (CaCE)cited by 135
  2. 2019Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systemscited by 130
  3. 2019CXPlain: Causal Explanations for Model Interpretation under Uncertaintycited by 98
  4. 2020Generative causal explanations of black-box classifierscited by 41
  5. 2018Explaining Deep Learning Models using Causal Inferencecited by 18
  6. 2019Explaining Visual Models by Causal Attributioncited by 18
  7. 2021Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglementcited by 12
  8. 2019Learning Interpretable Models with Causal Guaranteescited by 11
  9. 2020PatchVAE: Learning Local Latent Codes for Recognitioncited by 10
  10. 2020Discond-VAE: Disentangling Continuous Factors from the Discretecited by 4
  11. 2019Improving Disentangled Representation Learning with the Beta Bernoulli Processcited by 2

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2022: 2; 2023: 3; 2024: 2; 2025: 0; 2026: 1.

8 papers.

  1. 2023Causal Reinforcement Learning: A Surveycited by 22
  2. 2022ECLAD: Extracting Concepts with Local Aggregated Descriptorscited by 15
  3. 2022Deep Causal Learning: Representation, Discovery and Inferencecited by 11
  4. 2024A survey on Concept-based Approaches For Model Improvementcited by 8
  5. 2023Statistically Significant Concept-based Explanation of Image Classifiers via Model Knockoffscited by 4
  6. 2023SurroCBM: Concept Bottleneck Surrogate Models for Generative Post-hoc Explanationcited by 2
  7. 2024LatentExplainer: Explaining Latent Representations in Deep Generative Models with Multimodal Large Language Modelscited by 1
  8. 2026Towards Verified and Targeted Explanations through Formal Methodscited by 0