Interpreting and improving deep-learning models with reality checks
This work addresses the interpretability challenge in deep learning for researchers and practitioners, but it appears incremental as it builds on existing attribution methods.
The paper tackles the problem of interpreting deep-learning models by attributing importance to features and their interactions for individual predictions, and demonstrates how these attributions can improve model generalization or enable distillation into simpler models.
Recent deep-learning models have achieved impressive predictive performance by learning complex functions of many variables, often at the cost of interpretability. This chapter covers recent work aiming to interpret models by attributing importance to features and feature groups for a single prediction. Importantly, the proposed attributions assign importance to interactions between features, in addition to features in isolation. These attributions are shown to yield insights across real-world domains, including bio-imaging, cosmology image and natural-language processing. We then show how these attributions can be used to directly improve the generalization of a neural network or to distill it into a simple model. Throughout the chapter, we emphasize the use of reality checks to scrutinize the proposed interpretation techniques.