Explaining Deep Learning Models using Causal Inference
arXiv:1811.04376
Abstract
Although deep learning models have been successfully applied to a variety of tasks, due to the millions of parameters, they are becoming increasingly opaque and complex. In order to establish trust for their widespread commercial use, it is important to formalize a principled framework to reason over these models. In this work, we use ideas from causal inference to describe a general framework to reason over CNN models. Specifically, we build a Structural Causal Model (SCM) as an abstraction over a specific aspect of the CNN. We also formulate a method to quantitatively rank the filters of a convolution layer according to their counterfactual importance. We illustrate our approach with popular CNN architectures such as LeNet5, VGG19, and ResNet32.
References in corpus (6)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Counterfactual Fairness
- Gender Bias in Neural Natural Language Processing
- Causal Learning and Explanation of Deep Neural Networks via Autoencoded Activations
- Full deep neural network training on a pruned weight budget
- A Primer on Causal Analysis
Cited by in corpus (13)
- Causality Learning: A New Perspective for Interpretable Machine Learning
- Evaluating Explanation Without Ground Truth in Interpretable Machine Learning
- Socially Responsible AI Algorithms: Issues, Purposes, and Challenges
- Generative Counterfactual Introspection for Explainable Deep Learning
- When Causal Intervention Meets Adversarial Examples and Image Masking for Deep Neural Networks
- Generative causal explanations of black-box classifiers
- Causal Interpretability for Machine Learning -- Problems, Methods and Evaluation
- Adversarial Attacks and Defenses: An Interpretation Perspective
- Explainable Deep Learning for Uncovering Actionable Scientific Insights for Materials Discovery and Design
- Actionable Attribution Maps for Scientific Machine Learning
- Causal Learning for Socially Responsible AI
- Examining the causal structures of deep neural networks using information theory
- Causal Explanations of Image Misclassifications