Towards a Unified Evaluation of Explanation Methods without Ground Truth
arXiv:1911.09017
Abstract
This paper proposes a set of criteria to evaluate the objectiveness of explanation methods of neural networks, which is crucial for the development of explainable AI, but it also presents significant challenges. The core challenge is that people usually cannot obtain ground-truth explanations of the neural network. To this end, we design four metrics to evaluate explanation results without ground-truth explanations. Our metrics can be broadly applied to nine benchmark methods of interpreting neural networks, which provides new insights of explanation methods.
References in corpus (12)
- Striving for Simplicity: The All Convolutional Net
- Understanding Neural Networks Through Deep Visualization
- Understanding Black-box Predictions via Influence Functions
- InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
- Sanity Checks for Saliency Maps
- Not Just a Black Box: Learning Important Features Through Propagating Activation Differences
- Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)
- Learning how to explain neural networks: PatternNet and PatternAttribution
- Harnessing Deep Neural Networks with Logic Rules
- Benchmarking Attribution Methods with Relative Feature Importance
- Evaluating Explanation Without Ground Truth in Interpretable Machine Learning
- Can I Trust the Explainer? Verifying Post-hoc Explanatory Methods