Towards Stable and Structured Time Series Generation with Perturbation-Aware Flow Matching

TL;DR

This paper introduces PAFM, a perturbation-aware flow matching framework that improves the stability and structural consistency of time series generation by modeling localized disturbances and trajectory deviations.

Contribution

The paper proposes a novel perturbation-aware flow matching method with dual-path velocity fields and a mixture-of-experts decoder for better time series generation under perturbations.

Findings

Outperforms strong baselines in various generation tasks

Enhances structural coherence in perturbed time series

Demonstrates robustness to localized disturbances

Abstract

Time series generation is critical for a wide range of applications, which greatly supports downstream analytical and decision-making tasks. However, the inherent temporal heterogeneous induced by localized perturbations present significant challenges for generating structurally consistent time series. While flow matching provides a promising paradigm by modeling temporal dynamics through trajectory-level supervision, it fails to adequately capture abrupt transitions in perturbed time series, as the use of globally shared parameters constrains the velocity field to a unified representation. To address these limitations, we introduce \textbf{PAFM}, a \textbf{P}erturbation-\textbf{A}ware \textbf{F}low \textbf{M}atching framework that models perturbed trajectories to ensure stable and structurally consistent time series generation. The framework incorporates perturbation-guided training to simulate localized disturbances and leverages a dual-path velocity field to capture trajectory deviations under perturbation, enabling refined modeling of perturbed behavior to enhance the structural coherence. In order to further improve sensitivity to trajectory perturbations while enhancing expressiveness, a mixture-of-experts decoder with flow routing dynamically allocates modeling capacity in response to different trajectory dynamics. Extensive experiments on both unconditional and conditional generation tasks demonstrate that PAFM consistently outperforms strong baselines. Code is available at https://anonymous.4open.science/r/PAFM-03B2.