Introducing the Attribution Stability Indicator: a Measure for Time Series XAI Attributions

TL;DR

This paper introduces the Attribution Stability Indicator (ASI), a new measure for evaluating the robustness and trustworthiness of attribution methods in time series models, addressing a key challenge in interpretability.

Contribution

The paper proposes the ASI, extending perturbation analysis with correlation measures to evaluate attribution stability and trustworthiness in time series data.

Findings

ASI effectively captures attribution robustness.

Analysis shows ASI scores vary meaningfully across datasets.

The method enhances interpretability assessment in time series models.

Abstract

Given the increasing amount and general complexity of time series data in domains such as finance, weather forecasting, and healthcare, there is a growing need for state-of-the-art performance models that can provide interpretable insights into underlying patterns and relationships. Attribution techniques enable the extraction of explanations from time series models to gain insights but are hard to evaluate for their robustness and trustworthiness. We propose the Attribution Stability Indicator (ASI), a measure to incorporate robustness and trustworthiness as properties of attribution techniques for time series into account. We extend a perturbation analysis with correlations of the original time series to the perturbed instance and the attributions to include wanted properties in the measure. We demonstrate the wanted properties based on an analysis of the attributions in a dimension-reduced space and the ASI scores distribution over three whole time series classification datasets.