Interpreting Outliers in Time Series Data through Decoding Autoencoder

TL;DR

The paper addresses the challenge of interpreting outliers in manufacturing time series when using opaque autoencoder models. It applies XAI techniques to the encoder of a 1D CAE to generate latent-feature heatmaps and introduces Aggregated Explanatory Ensemble (AEE) to fuse Grad-CAM, LIME, SHAP, and LRP explanations into a single, more informative view, complemented by a quality measure QM_e that evaluates explanation impact in latent space. The approach is validated through a large industrial dataset, combining qualitative domain-expert assessment with quantitative QM_e analysis, and demonstrates improved interpretability of anomalies without sacrificing detection performance. The work advances practical explainability in safety-critical manufacturing settings by providing actionable, encoder-centered insights and a systematic method to compare explanations.

Abstract

Outlier detection is a crucial analytical tool in various fields. In critical systems like manufacturing, malfunctioning outlier detection can be costly and safety-critical. Therefore, there is a significant need for explainable artificial intelligence (XAI) when deploying opaque models in such environments. This study focuses on manufacturing time series data from a German automotive supply industry. We utilize autoencoders to compress the entire time series and then apply anomaly detection techniques to its latent features. For outlier interpretation, we (i) adopt widely used XAI techniques to the autoencoder's encoder. Additionally, (ii) we propose AEE, Aggregated Explanatory Ensemble, a novel approach that fuses explanations of multiple XAI techniques into a single, more expressive interpretation. For evaluation of explanations, (iii) we propose a technique to measure the quality of encoder explanations quantitatively. Furthermore, we qualitatively assess the effectiveness of outlier explanations with domain expertise.

Figures (8)

• Figure 1: Aggregated Explanatory Ensemble - AEE. The aggregated explanation is represented by a heat map in the background, with deeper shades of red indicating areas of greater significance for its explanation in the time series. The black curve visualizes an anomalous time series, with the explanation highlighting a disruption in the pattern between the 6300 and 7200 marks in the time series.
• Figure 2: Individual XAI Results. The XAI results are presented in the form of heatmaps. The black portions of the images denote the time series signal. These displayed instances were identified as abnormal by the AE's pipeline. The heatmap in the background indicates feature importance using varying intensities of red. We must evaluate color intensity individually as XAI techniques calculate feature importance differently.
• Figure 3: AEE XAI Results. This figure presents the results of the XAI analysis for the AEE approach. The format and layout of these explanations are consistent with those shown in \ref{['fig:main_1']}.
• Figure 4: Interquartile Range Quality Measurements. The visualization depicts quality measurement scores for each XAI technique, categorized into true anomalies (NOK) and false anomalies (OK). The measurements are further stratified into noise ($\mathbf{t^c_r}$ - XAI shuffled), denoted by green, and XAI ($\mathbf{t^c}$ - XAI perturbated), represented by red.
• Figure 5: Grad-CAM - Individual Feature Heatmaps. The images illustrate individual latent feature explanations in the form of a heatmap generated by Grad-CAM. The black curve illustrates the original time series, while the red curve represents its reconstruction.