GenIAS: Generator for Instantiating Anomalies in time Series

2025-02-12

AI Summary

Existing time-series anomaly detection (TSAD) methods rely on hand-crafted anomaly injection, failing to capture the natural diversity of real-world anomalies and often restricting evaluation to univariate settings. To address these limitations, this paper proposes the first VAE-based generative framework for synthesizing anomalies in multivariate time series. It introduces a novel latent-space learning perturbation mechanism, coupled with a triplet loss incorporating min-max margin constraints and variance scaling, enabling cross-scale, multi-segment, and controllable anomaly injection. The method preserves anomaly realism while significantly enhancing diversity and inherently supports unified modeling of both univariate and multivariate sequences. Extensive experiments on standard TSAD benchmarks demonstrate consistent and substantial improvements over 17 state-of-the-art baselines, validating its effectiveness in boosting detection performance.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• 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: Web data generation and simulation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

A recent and promising approach for building time series anomaly detection (TSAD) models is to inject synthetic samples of anomalies within real data sets. The existing injection mechanisms have significant limitations - most of them rely on ad hoc, hand-crafted strategies which fail to capture the natural diversity of anomalous patterns, or are restricted to univariate time series settings. To address these challenges, we design a generative model for TSAD using a variational autoencoder, which is referred to as a Generator for Instantiating Anomalies in Time Series (GenIAS). GenIAS is designed to produce diverse and realistic synthetic anomalies for TSAD tasks. By employing a novel learned perturbation mechanism in the latent space and injecting the perturbed patterns in different segments of time series, GenIAS can generate anomalies with greater diversity and varying scales. Further, guided by a new triplet loss function, which uses a min-max margin and a new variance-scaling approach to further enforce the learning of compact normal patterns, GenIAS ensures that anomalies are distinct from normal samples while remaining realistic. The approach is effective for both univariate and multivariate time series. We demonstrate the diversity and realism of the generated anomalies. Our extensive experiments demonstrate that GenIAS - when integrated into a TSAD task - consistently outperforms seventeen traditional and deep anomaly detection models, thereby highlighting the potential of generative models for time series anomaly generation.

Problem

• Generates diverse synthetic time series anomalies
• Addresses limitations in current anomaly injection methods
• Improves anomaly detection in univariate and multivariate series

Innovation

• Variational autoencoder for anomaly generation
• Learned perturbation mechanism in latent space
• Triplet loss function for pattern distinction