Adversarial Framework with Certified Robustness for Time-Series Domain via Statistical Features

Time-series data arises in many real-world applications (e.g., mobile health) and deep neural networks (DNNs) have shown great success solving them. Despite their success, little is known about robustness to adversarial attacks. In this paper, we propose a novel framework referred as Time-Series Attacks via STATistical Features (TSA-STAT). To address the unique challenges of time-series domain, TSA-STAT employs constraints on statistical features construct examples. Optimized polynomial transformations are used create attacks that more effective (in terms successfully fooling DNNs) than those based additive perturbations. We also provide certified bounds norm for constructing Our experiments diverse benchmark datasets show effectiveness DNNs domain improving robustness.