DATA-ADAPTIVE DISCRIMINATIVE FEATURE LOCALIZATION WITH STATISTICALLY GUARANTEED INTERPRETATION

• 成果类型: Article
• 署名作者: Dai, Ben; Shen, Xiaotong; Chen, Lin yee; Li, Chunlin; Pan, Wei
• 署名单位: Chinese University of Hong Kong; University of Minnesota System; University of Minnesota Twin Cities; University of Minnesota System; University of Minnesota Twin Cities; University of Minnesota System; University of Minnesota Twin Cities; University of Minnesota System; University of Minnesota Twin Cities
• 刊物名称: ANNALS OF APPLIED STATISTICS
• ISSN/ISSBN: 1932-6157
• DOI: 10.1214/22-AOAS1705
• 发表日期: 2023
• 页码: 2019-2038
• 关键词: Classification

摘要

In explainable artificial intelligence, discriminative feature localization is critical to reveal a black-box 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 R2. We apply the proposed method to the MNIST dataset and the MIT-BIH dataset with a convolutional autoencoder. 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 that implements the proposed approach.