On the Consistency and Robustness of Saliency Explanations for Time Series Classification

TL;DR

This paper critically analyzes the consistency and robustness of saliency map explanations for time series classification, revealing their limitations and motivating the development of improved interpretability methods.

Contribution

The study provides an extensive evaluation of saliency explanations for time series, highlighting their shortcomings and emphasizing the need for more reliable interpretability techniques.

Findings

Saliency maps lack consistency across different explanations.

Saliency explanations are not robust to data perturbations.

Current methods show limited reliability for time series interpretation.

Abstract

Interpretable machine learning and explainable artificial intelligence have become essential in many applications. The trade-off between interpretability and model performance is the traitor to developing intrinsic and model-agnostic interpretation methods. Although model explanation approaches have achieved significant success in vision and natural language domains, explaining time series remains challenging. The complex pattern in the feature domain, coupled with the additional temporal dimension, hinders efficient interpretation. Saliency maps have been applied to interpret time series windows as images. However, they are not naturally designed for sequential data, thus suffering various issues. This paper extensively analyzes the consistency and robustness of saliency maps for time series features and temporal attribution. Specifically, we examine saliency explanations from both perturbation-based and gradient-based explanation models in a time series classification task. Our experimental results on five real-world datasets show that they all lack consistent and robust performances to some extent. By drawing attention to the flawed saliency explanation models, we motivate to develop consistent and robust explanations for time series classification.