xLSTMAD: A Powerful xLSTM-based Method for Anomaly Detection

2025-06-28

AI Summary

This work addresses the lack of models capable of effective long-range dependency modeling in multivariate time series anomaly detection. We introduce xLSTMAD—the first anomaly detection method built upon a full encoder-decoder xLSTM architecture—marking the first application of xLSTM to this task. Its key innovations are: (1) a dual-branch decoder supporting both future-value forecasting (xLSTMAD-F) and input reconstruction (xLSTMAD-R) within a unified framework; and (2) integration of multiplicative gating, residual connections, and a joint SoftDTW–MSE loss to jointly optimize local fidelity and global temporal alignment. Evaluated on the TSB-AD-M benchmark, xLSTMAD significantly outperforms 23 state-of-the-art baselines, achieving new SOTA performance on the VUS-PR metric. These results demonstrate the effectiveness and generalization capability of xLSTM for time series anomaly detection.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The recently proposed xLSTM is a powerful model that leverages expressive multiplicative gating and residual connections, providing the temporal capacity needed for long-horizon forecasting and representation learning. This architecture has demonstrated success in time series forecasting, lossless compression, and even large-scale language modeling tasks, where its linear memory footprint and fast inference make it a viable alternative to Transformers. Despite its growing popularity, no prior work has explored xLSTM for anomaly detection. In this work, we fill this gap by proposing xLSTMAD, the first anomaly detection method that integrates a full encoder-decoder xLSTM architecture, purpose-built for multivariate time series data. Our encoder processes input sequences to capture historical context, while the decoder is devised in two separate variants of the method. In the forecasting approach, the decoder iteratively generates forecasted future values xLSTMAD-F, while the reconstruction approach reconstructs the input time series from its encoded counterpart xLSTMAD-R. We investigate the performance of two loss functions: Mean Squared Error (MSE), and Soft Dynamic Time Warping (SoftDTW) to consider local reconstruction fidelity and global sequence alignment, respectively. We evaluate our method on the comprehensive TSB-AD-M benchmark, which spans 17 real-world datasets, using state-of-the-art challenging metrics such as VUS-PR. In our results, xLSTM showcases state-of-the-art accuracy, outperforming 23 popular anomaly detection baselines. Our paper is the first work revealing the powerful modeling capabilities of xLSTM for anomaly detection, paving the way for exciting new developments on this subject. Our code is available at: https://github.com/Nyderx/xlstmad

Problem

• Proposes xLSTMAD for anomaly detection in multivariate time series
• First method using full encoder-decoder xLSTM architecture
• Evaluates performance on 17 datasets outperforming 23 baselines

Innovation

• Uses xLSTM encoder-decoder for anomaly detection
• Integrates forecasting and reconstruction variants
• Employs MSE and SoftDTW loss functions