STAR: Boosting Time Series Foundation Models for Anomaly Detection through State-aware Adapter

TL;DR

STAR tackles the challenge of leveraging discrete state variables in time series anomaly detection by introducing a state-aware adapter that redefines how state information is processed during fine-tuning. It combines an Identity-guided State Encoder with a State Memory, a Conditional Bottleneck Adapter with state-conditioned low-rank updates, and a Numeral-State Matching mechanism to align numeral and state representations. Across five real-world datasets and multiple backbones, STAR yields consistent improvements in anomaly detection performance, with notable gains on datasets where state dynamics strongly influence numerical behavior. The approach is plug-and-play, maintains backbone parameters during fine-tuning, and offers practical benefits for industrial deployment where state information is rich and essential for accurate detection.

Abstract

While Time Series Foundation Models (TSFMs) have demonstrated remarkable success in Multivariate Time Series Anomaly Detection (MTSAD), however, in real-world industrial scenarios, many time series comprise not only numerical variables such as temperature and flow, but also numerous discrete state variables that describe the system status, such as valve on/off or day of the week. Existing TSFMs often overlook the distinct categorical nature of state variables and their critical role as conditions, typically treating them uniformly with numerical variables. This inappropriate modeling approach prevents the model from fully leveraging state information and even leads to a significant degradation in detection performance after state variables are integrated. To address this critical limitation, this paper proposes a novel STate-aware AdapteR (STAR). STAR is a plug-and-play module designed to enhance the capability of TSFMs in modeling and leveraging state variables during the fine-tuning stage. Specifically, STAR comprisesthree core components: (1) We design an Identity-guided State Encoder, whicheffectively captures the complex categorical semantics of state variables through a learnable State Memory. (2) We propose a Conditional Bottleneck Adapter, which dynamically generates low-rank adaptation parameters conditioned on the current state, thereby flexibly injecting the influence of state variables into the backbone model. (3) We also introduce a Numeral-State Matching module to more effectively detect anomalies inherent to the state variables themselves. Extensive experiments conducted on real-world datasets demonstrate that STAR can improve the performance of existing TSFMs on MTSAD.

Figures (8)

• Figure 1: (a) Case of State variables (Valve) and numeral variables (Water Flow) in SWaT. (b) State variables are prevalent in real-world datasets and frequently constitute a substantial proportion. (c) Impact of incorporating state variables on detection performance.
• Figure 2: The overview of STAR.
• Figure 3: Parameter sensitivity studies of main hyper-parameters in STAR.
• Figure 4: (a) shows the similarity between the embeddings output from Identity-guided State Encoder. (b) shows the weight of selection in Memory Router.
• Figure 5: Case series in Genesis.