Data-Adaptive Discriminative Feature Localization with Statistically Guaranteed Interpretation
This work addresses the need for interpretable AI by providing a statistically guaranteed method for feature localization, which is incremental as it builds on adversarial attack techniques but adds formal guarantees.
The paper tackles the problem of discriminative feature localization in explainable AI by developing a framework based on adversarial attacks, which localizes features with statistical guarantees using a generalized partial R² measure, and applies it to MNIST and MIT-BIH datasets, showing visually appealing and biologically plausible results that compare favorably with state-of-the-art methods.
In explainable artificial intelligence, discriminative feature localization is critical to reveal a blackbox model's decision-making process from raw data to prediction. In this article, we use two real datasets, the MNIST handwritten digits and MIT-BIH Electrocardiogram (ECG) signals, to motivate key characteristics of discriminative features, namely adaptiveness, predictive importance and effectiveness. Then, we develop a localization framework based on adversarial attacks to effectively localize discriminative features. In contrast to existing heuristic methods, we also provide a statistically guaranteed interpretability of the localized features by measuring a generalized partial $R^2$. We apply the proposed method to the MNIST dataset and the MIT-BIH dataset with a convolutional auto-encoder. In the first, the compact image regions localized by the proposed method are visually appealing. Similarly, in the second, the identified ECG features are biologically plausible and consistent with cardiac electrophysiological principles while locating subtle anomalies in a QRS complex that may not be discernible by the naked eye. Overall, the proposed method compares favorably with state-of-the-art competitors. Accompanying this paper is a Python library dnn-locate (https://dnn-locate.readthedocs.io/en/latest/) that implements the proposed approach.