Explaining Visual Models by Causal Attribution
arXiv:1909.08891
Abstract
Model explanations based on pure observational data cannot compute the effects of features reliably, due to their inability to estimate how each factor alteration could affect the rest. We argue that explanations should be based on the causal model of the data and the derived intervened causal models, that represent the data distribution subject to interventions. With these models, we can compute counterfactuals, new samples that will inform us how the model reacts to feature changes on our input. We propose a novel explanation methodology based on Causal Counterfactuals and identify the limitations of current Image Generative Models in their application to counterfactual creation.
arXiv comment2019 ICCV Workshop on Interpreting and Explaining Visual Artificial Intelligence Models
References in corpus (6)
- Auto-Encoding Variational Bayes
- Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
- Implicit Reparameterization Gradients
- Explaining Image Classifiers by Counterfactual Generation
- Gradients of Counterfactuals
- Image Counterfactual Sensitivity Analysis for Detecting Unintended Bias
Cited by in corpus (5)
- Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers
- Socially Responsible AI Algorithms: Issues, Purposes, and Challenges
- Generative causal explanations of black-box classifiers
- Causal Interpretability for Machine Learning -- Problems, Methods and Evaluation
- VACA: Design of Variational Graph Autoencoders for Interventional and Counterfactual Queries