Generative Counterfactual Introspection for Explainable Deep Learning
This work addresses the need for explainable AI in deep learning, though it appears incremental as it builds on existing generative and counterfactual methods.
The authors tackled the problem of interpreting deep neural networks by developing a generative model-based introspection technique that edits input images to answer counterfactual questions, demonstrating its application on MNIST and CelebA datasets.
In this work, we propose an introspection technique for deep neural networks that relies on a generative model to instigate salient editing of the input image for model interpretation. Such modification provides the fundamental interventional operation that allows us to obtain answers to counterfactual inquiries, i.e., what meaningful change can be made to the input image in order to alter the prediction. We demonstrate how to reveal interesting properties of the given classifiers by utilizing the proposed introspection approach on both the MNIST and the CelebA dataset.