Learning Perturbations to Explain Time Series Predictions

TL;DR

This paper introduces a novel approach for explaining multivariate time series predictions by learning both masks and perturbations, leading to more accurate and meaningful explanations of feature impacts over time.

Contribution

It proposes a new method that learns perturbations alongside masks to improve the interpretability of time series prediction explanations.

Findings

Learning perturbations enhances explanation quality

Method outperforms fixed perturbation approaches

Provides more accurate feature-time impact insights

Abstract

Explaining predictions based on multivariate time series data carries the additional difficulty of handling not only multiple features, but also time dependencies. It matters not only what happened, but also when, and the same feature could have a very different impact on a prediction depending on this time information. Previous work has used perturbation-based saliency methods to tackle this issue, perturbing an input using a trainable mask to discover which features at which times are driving the predictions. However these methods introduce fixed perturbations, inspired from similar methods on static data, while there seems to be little motivation to do so on temporal data. In this work, we aim to explain predictions by learning not only masks, but also associated perturbations. We empirically show that learning these perturbations significantly improves the quality of these explanations on time series data.