CAST: Context- and Anomaly Structure-Conditioned Time Series Anomaly Generation
cs.LG
Submitted: 2026-08-19
Updated: 2026-08-19
Code: https://github.com/UNITES-Lab/FlowTS
License: http://creativecommons.org/licenses/by/4.0/
The gist: Anomalous time series play a critical role in safety-critical domains, yet they are inherently scarce, heterogeneous, and costly to obtain.
Terminology
Abstract
Anomalous time series play a critical role in safety-critical domains, yet they are inherently scarce, heterogeneous, and costly to obtain. Existing time series generation methods predominantly focus on synthesizing normal data, providing limited value when anomalous samples are needed. We identify two fundamental challenges in anomaly generation: (i) the scarcity of anomaly data, and (ii) the heterogeneous morphological characteristics of anomalies. To address these challenges, we propose CAST, a Context- and Anomaly Structure-conditioned Time series anomaly generation framework with principled two-stage pretraining and finetuning strategy. In pretraining stage, we leverage abundant normal time series data to learn underlying system dynamics and substantially mitigate the limited availability of anomaly data. During finetuning, CAST explicitly conditions the generator on learned anomaly structure representations, enabling it to capture heterogeneous anomaly morphologies under similar contextual conditions. Extensive experiments on multiple real-world univariate and multivariate datasets demonstrate that CAST consistently outperforms state-of-the-art anomaly generation methods in terms of both generation fidelity and downstream task utility, highlighting the effectiveness of the proposed approach.
Sources
- Neural Contextual Anomaly Detection for Time Series
- LTSM-Bundle: A Toolbox and Benchmark on Large Language Models for Time Series Forecasting
- GenIAS: Generator for Instantiating Anomalies in time Series
- TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generation
- A Diffusion Model for Regular Time Series Generation from Irregular Data with Completion and Masking
- RobustTAD: Robust Time Series Anomaly Detection via Decomposition and Convolutional Neural Networks
- MOMENT: A Family of Open Time-series Foundation Models
- FlowTS: Time Series Generation via Rectified Flow
- Robust PCA for Anomaly Detection in Cyber Networks
- Training-Free Time Series Classification via In-Context Reasoning with LLM Agents
- Unsupervised Representation Learning for Time Series with Temporal Neighborhood Coding
- TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis
- mixup: Beyond Empirical Risk Minimization
- Towards Stable and Structured Time Series Generation with Perturbation-Aware Flow Matching
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks