Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cho, Dongchan, Han, Jiho, Kang, Keumyeong, Kim, Minsang, Ryu, Honggyu, Jung, Namsoon
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914102788489216
author Cho, Dongchan
Han, Jiho
Kang, Keumyeong
Kim, Minsang
Ryu, Honggyu
Jung, Namsoon
author_facet Cho, Dongchan
Han, Jiho
Kang, Keumyeong
Kim, Minsang
Ryu, Honggyu
Jung, Namsoon
contents Real-world multivariate time series anomalies are rare and often unlabeled. Additionally, prevailing methods rely on increasingly complex architectures tuned to benchmarks, detecting only fragments of anomalous segments and overstating performance. In this paper, we introduce OracleAD, a simple and interpretable unsupervised framework for multivariate time series anomaly detection. OracleAD encodes each variable's past sequence into a single causal embedding to jointly predict the present time point and reconstruct the input window, effectively modeling temporal dynamics. These embeddings then undergo a self-attention mechanism to project them into a shared latent space and capture spatial relationships. These relationships are not static, since they are modeled by a property that emerges from each variable's temporal dynamics. The projected embeddings are aligned to a Stable Latent Structure (SLS) representing normal-state relationships. Anomalies are identified using a dual scoring mechanism based on prediction error and deviation from the SLS, enabling fine-grained anomaly diagnosis at each time point and across individual variables. Since any noticeable SLS deviation originates from embeddings that violate the learned temporal causality of normal data, OracleAD directly pinpoints the root-cause variables at the embedding level. OracleAD achieves state-of-the-art results across multiple real-world datasets and evaluation protocols, while remaining interpretable through SLS.
format Preprint
id arxiv_https___arxiv_org_abs_2510_16511
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection
Cho, Dongchan
Han, Jiho
Kang, Keumyeong
Kim, Minsang
Ryu, Honggyu
Jung, Namsoon
Machine Learning
Artificial Intelligence
Real-world multivariate time series anomalies are rare and often unlabeled. Additionally, prevailing methods rely on increasingly complex architectures tuned to benchmarks, detecting only fragments of anomalous segments and overstating performance. In this paper, we introduce OracleAD, a simple and interpretable unsupervised framework for multivariate time series anomaly detection. OracleAD encodes each variable's past sequence into a single causal embedding to jointly predict the present time point and reconstruct the input window, effectively modeling temporal dynamics. These embeddings then undergo a self-attention mechanism to project them into a shared latent space and capture spatial relationships. These relationships are not static, since they are modeled by a property that emerges from each variable's temporal dynamics. The projected embeddings are aligned to a Stable Latent Structure (SLS) representing normal-state relationships. Anomalies are identified using a dual scoring mechanism based on prediction error and deviation from the SLS, enabling fine-grained anomaly diagnosis at each time point and across individual variables. Since any noticeable SLS deviation originates from embeddings that violate the learned temporal causality of normal data, OracleAD directly pinpoints the root-cause variables at the embedding level. OracleAD achieves state-of-the-art results across multiple real-world datasets and evaluation protocols, while remaining interpretable through SLS.
title Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2510.16511