CHIME: Conditional Hallucination and Integrated Multi-scale Enhancement for Time Series Diffusion Model

TL;DR

CHIME is a novel framework for time series diffusion models that enhances multi-scale feature alignment and long-term feature transfer, leading to improved generation quality and generalization in real-world datasets.

Contribution

It introduces multi-scale decomposition and a feature hallucination module to improve time series generation across different scales and entities.

Findings

Achieves state-of-the-art performance on real-world datasets.

Exhibits excellent generalization in few-shot scenarios.

Enhances long-time scale feature transfer.

Abstract

The denoising diffusion probabilistic model has become a mainstream generative model, achieving significant success in various computer vision tasks. Recently, there has been initial exploration of applying diffusion models to time series tasks. However, existing studies still face challenges in multi-scale feature alignment and generative capabilities across different entities and long-time scales. In this paper, we propose CHIME, a conditional hallucination and integrated multi-scale enhancement framework for time series diffusion models. By employing multi-scale decomposition and integration, CHIME captures the decomposed features of time series, achieving in-domain distribution alignment between generated and original samples. In addition, we introduce a feature hallucination module in the conditional denoising process, enabling the temporal features transfer across long-time scales. Experimental results on publicly available real-world datasets demonstrate that CHIME achieves state-of-the-art performance and exhibits excellent generative generalization capabilities in few-shot scenarios.