Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
139 tokens/sec
GPT-4o
47 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Graph Edits for Counterfactual Explanations: A comparative study (2401.11609v3)

Published 21 Jan 2024 in cs.LG and cs.AI

Abstract: Counterfactuals have been established as a popular explainability technique which leverages a set of minimal edits to alter the prediction of a classifier. When considering conceptual counterfactuals on images, the edits requested should correspond to salient concepts present in the input data. At the same time, conceptual distances are defined by knowledge graphs, ensuring the optimality of conceptual edits. In this work, we extend previous endeavors on graph edits as counterfactual explanations by conducting a comparative study which encompasses both supervised and unsupervised Graph Neural Network (GNN) approaches. To this end, we pose the following significant research question: should we represent input data as graphs, which is the optimal GNN approach in terms of performance and time efficiency to generate minimal and meaningful counterfactual explanations for black-box image classifiers?

Definition Search Book Streamline Icon: https://streamlinehq.com
References (43)
  1. Meaningfully Debugging Model Mistakes using Conceptual Counterfactual Explanations. arXiv:2106.12723.
  2. CoCoX: Generating Conceptual and Counterfactual Explanations via Fault-Lines. Proceedings of the AAAI Conference on Artificial Intelligence, 34: 2594–2601.
  3. Diffusion Visual Counterfactual Explanations. arXiv:2210.11841.
  4. Natural Language Processing with Python: Analyzing Text with the Natural Language Toolkit. Beijing: O’Reilly. ISBN 978-0-596-51649-9.
  5. Shortest-path kernels on graphs. In Fifth IEEE International Conference on Data Mining (ICDM’05), 8 pp.–.
  6. Semantics and explanation: why counterfactual explanations produce adversarial examples in deep neural networks. arXiv:2012.10076.
  7. Explaining Image Classifiers by Counterfactual Generation. arXiv:1807.08024.
  8. A Comprehensive Survey of Scene Graphs: Generation and Application. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(1): 1–26.
  9. Choose your Data Wisely: A Framework for Semantic Counterfactuals. arXiv:2305.17667.
  10. Latent Diffusion Counterfactual Explanations. ArXiv, abs/2310.06668.
  11. Fast Graph Representation Learning with PyTorch Geometric. In ICLR Workshop on Representation Learning on Graphs and Manifolds.
  12. Conceptual Edits as Counterfactual Explanations. In Proceedings of the AAAI 2022 Spring Symposium on Machine Learning and Knowledge Engineering for Hybrid Intelligence (AAAI-MAKE 2022), Stanford University, Palo Alto, California, USA.
  13. On Graph Kernels: Hardness Results and Efficient Alternatives. In Schölkopf, B.; and Warmuth, M. K., eds., Learning Theory and Kernel Machines, 129–143. Berlin, Heidelberg: Springer Berlin Heidelberg. ISBN 978-3-540-45167-9.
  14. Generative Adversarial Nets. In Ghahramani, Z.; Welling, M.; Cortes, C.; Lawrence, N.; and Weinberger, K., eds., Advances in Neural Information Processing Systems, volume 27. Curran Associates, Inc.
  15. Counterfactual Visual Explanations. arXiv:1904.07451.
  16. A Graph Feature Auto-Encoder for the Prediction of Unobserved Node Features on Biological Networks. arXiv:2005.03961.
  17. Grounding Visual Explanations. arXiv:1807.09685.
  18. A Linear-Time Graph Kernel. In 2009 Ninth IEEE International Conference on Data Mining, 179–188.
  19. A shortest augmenting path algorithm for dense and sparse linear assignment problems. Computing, 38(4): 325–340.
  20. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907.
  21. Variational Graph Auto-Encoders. arXiv:1611.07308.
  22. Kondor, R. 2002. Diffusion kernels on graphs and other discrete structures. In International Conference on Machine Learning.
  23. Visual genome: Connecting language and vision using crowdsourced dense image annotations. International journal of computer vision, 123(1): 32–73.
  24. Graph Matching Networks for Learning the Similarity of Graph Structured Objects. arXiv:1904.12787.
  25. Decoupled Weight Decay Regularization. arXiv:1711.05101.
  26. Counterfactual Edits for Generative Evaluation. arXiv:2303.01555.
  27. Introduction to Information Retrieval. USA: Cambridge University Press. ISBN 0521865719.
  28. Miller, G. A. 1995. WordNet: a lexical database for English. Communications of the ACM, 38(11): 39–41.
  29. OpenAI. 2023a. ChatGPT: Conversational Language Model. https://www.openai.com/research/chatgpt.
  30. OpenAI. 2023b. GPT-4 Technical Report. ArXiv, abs/2303.08774.
  31. Adversarially Regularized Graph Autoencoder for Graph Embedding. arXiv:1802.04407.
  32. FACE: Feasible and Actionable Counterfactual Explanations. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, AIES ’20. ACM.
  33. Pržulj, N. 2007. Biological network comparison using graphlet degree distribution. Bioinformatics, 23(2): e177–e183.
  34. Rudin, C. 2019. Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead. arXiv:1811.10154.
  35. A distance measure between attributed relational graphs for pattern recognition. IEEE Transactions on Systems, Man, and Cybernetics, SMC-13(3): 353–362.
  36. Weisfeiler-Lehman Graph Kernels. Journal of Machine Learning Research, 12(77): 2539–2561.
  37. GraKeL: A Graph Kernel Library in Python. arXiv:1806.02193.
  38. Making Heads or Tails: Towards Semantically Consistent Visual Counterfactuals. arXiv preprint arXiv:2203.12892.
  39. Graph attention networks. arXiv preprint arXiv:1710.10903.
  40. Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR. arXiv:1711.00399.
  41. How powerful are graph neural networks? arXiv preprint arXiv:1810.00826.
  42. Generating Natural Counterfactual Visual Explanations. In International Joint Conference on Artificial Intelligence.
  43. Places: A 10 million Image Database for Scene Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now