FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Runze, Wang, Hanchen, Zhang, Wenjie, Li, Binghao, Zhang, Yu, Lin, Xuemin, Zhang, Ying
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911324054749184
author Li, Runze
Wang, Hanchen
Zhang, Wenjie
Li, Binghao
Zhang, Yu
Lin, Xuemin
Zhang, Ying
author_facet Li, Runze
Wang, Hanchen
Zhang, Wenjie
Li, Binghao
Zhang, Yu
Lin, Xuemin
Zhang, Ying
contents Multivariate time series imputation is fundamental in applications such as healthcare, traffic forecasting, and biological modeling, where sensor failures and irregular sampling lead to pervasive missing values. However, existing Transformer- and diffusion-based models lack explicit inductive biases and frequency awareness, limiting their generalization under structured missing patterns and distribution shifts. We propose FADTI, a diffusion-based framework that injects frequency-informed feature modulation via a learnable Fourier Bias Projection (FBP) module and combines it with temporal modeling through self-attention and gated convolution. FBP supports multiple spectral bases, enabling adaptive encoding of both stationary and non-stationary patterns. This design injects frequency-domain inductive bias into the generative imputation process. Experiments on multiple benchmarks, including a newly introduced biological time series dataset, show that FADTI consistently outperforms state-of-the-art methods, particularly under high missing rates. Code is available at https://anonymous.4open.science/r/TimeSeriesImputation-52BF
format Preprint
id arxiv_https___arxiv_org_abs_2512_15116
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation
Li, Runze
Wang, Hanchen
Zhang, Wenjie
Li, Binghao
Zhang, Yu
Lin, Xuemin
Zhang, Ying
Machine Learning
Artificial Intelligence
Multivariate time series imputation is fundamental in applications such as healthcare, traffic forecasting, and biological modeling, where sensor failures and irregular sampling lead to pervasive missing values. However, existing Transformer- and diffusion-based models lack explicit inductive biases and frequency awareness, limiting their generalization under structured missing patterns and distribution shifts. We propose FADTI, a diffusion-based framework that injects frequency-informed feature modulation via a learnable Fourier Bias Projection (FBP) module and combines it with temporal modeling through self-attention and gated convolution. FBP supports multiple spectral bases, enabling adaptive encoding of both stationary and non-stationary patterns. This design injects frequency-domain inductive bias into the generative imputation process. Experiments on multiple benchmarks, including a newly introduced biological time series dataset, show that FADTI consistently outperforms state-of-the-art methods, particularly under high missing rates. Code is available at https://anonymous.4open.science/r/TimeSeriesImputation-52BF
title FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2512.15116