![[2601.21812] A Decomposable Forward Process in Diffusion Models for Time-Series Forecasting](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2601.21812] A Decomposable Forward Process in Diffusion Models for Time-Series Forecasting
Summary
This paper presents a novel forward diffusion process for time-series forecasting that effectively decomposes signals into spectral components, enhancing the preservation of temporal patterns like seasonality.
Why It Matters
The research addresses limitations in existing diffusion models for time-series forecasting by introducing a method that maintains high signal integrity. This advancement is crucial for improving forecast accuracy in various applications, such as finance and climate modeling, where understanding temporal patterns is essential.
Key Takeaways
- Introduces a model-agnostic forward diffusion process for time-series forecasting.
- Enhances preservation of temporal patterns by decomposing signals into spectral components.
- Maintains high signal-to-noise ratios for dominant frequencies, improving forecast quality.
- Compatible with existing diffusion models, requiring negligible computational overhead.
- Demonstrates improved performance across standard forecasting benchmarks.
Statistics > Machine Learning arXiv:2601.21812 (stat) [Submitted on 29 Jan 2026 (v1), last revised 16 Feb 2026 (this version, v2)] Title:A Decomposable Forward Process in Diffusion Models for Time-Series Forecasting Authors:Francisco Caldas, Sahil Kumar, Cláudia Soares View a PDF of the paper titled A Decomposable Forward Process in Diffusion Models for Time-Series Forecasting, by Francisco Caldas and 2 other authors View PDF HTML (experimental) Abstract:We introduce a model-agnostic forward diffusion process for time-series forecasting that decomposes signals into spectral components, preserving structured temporal patterns such as seasonality more effectively than standard diffusion. Unlike prior work that modifies the network architecture or diffuses directly in the frequency domain, our proposed method alters only the diffusion process itself, making it compatible with existing diffusion backbones (e.g., DiffWave, TimeGrad, CSDI). By staging noise injection according to component energy, it maintains high signal-to-noise ratios for dominant frequencies throughout the diffusion trajectory, thereby improving the recoverability of long-term patterns. This strategy enables the model to maintain the signal structure for a longer period in the forward process, leading to improved forecast quality. Across standard forecasting benchmarks, we show that applying spectral decomposition strategies, such as the Fourier or Wavelet transform, consistently improves upon diffusion mode...
