NormLime: A New Feature Importance Metric for Explaining Deep Neural Networks

TL;DR

NormLIME introduces a novel feature importance metric that effectively aggregates local explanations into accurate, class-specific global interpretations for deep neural networks, with strong human and numerical validation.

Contribution

It proposes NormLIME, a new method for aggregating local explanations into global, class-specific interpretations, improving interpretability of deep models.

Findings

Human user study favors NormLIME's class-specific explanations

Numerical experiments confirm NormLIME's effectiveness in identifying important features

NormLIME outperforms existing feature importance metrics

Abstract

The problem of explaining deep learning models, and model predictions generally, has attracted intensive interest recently. Many successful approaches forgo global approximations in order to provide more faithful local interpretations of the model's behavior. LIME develops multiple interpretable models, each approximating a large neural network on a small region of the data manifold and SP-LIME aggregates the local models to form a global interpretation. Extending this line of research, we propose a simple yet effective method, NormLIME for aggregating local models into global and class-specific interpretations. A human user study strongly favored class-specific interpretations created by NormLIME to other feature importance metrics. Numerical experiments confirm that NormLIME is effective at recognizing important features.