aiqbench analysis
FADTI advances multivariate time series imputation by integrating Fourier-based spectral guidance with attention mechanisms to better handle structured missing data in domains like healthcare and traffic forecasting. This approach addresses a key limitation in current diffusion models that rely solely on time-domain modeling, offering improved accuracy and sampling efficiency under high missing rates. The method represents a significant step forward in machine learning techniques for recovering complex temporal patterns from incomplete sensor data.
FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation
arXiv:2512.15116v3 Announce Type: replace-cross Abstract: 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. Existing Transformer- and diffusion-based imputers achieve strong performance, but they often rely mainly on time-domain modeling and lack adaptive spectral bias for recovering structured temporal gaps. We propose FADTI, a Fourier- and attention-driven diffusion framework for multivariate time series imputation. FADTI introduces a Fourier Bias Projection (FBP) module that injects learnable frequency-aware bias into intermediate hidden states during denoising. It projects intermediate hidden states onto Fourier bases, avoiding direct spectral estimation from masked or zero-filled inputs. With DFT, STFT, and FSST instantiations, FBP captures global periodicity, localized time--frequency variations, and non-stationary oscillatory patterns. By coupling FBP with self-attention and gated convolution, FADTI integrates frequency-domain guidance, temporal dependency modeling, and probabilistic denoising in a unified framework. Experiments on multiple benchmarks, including a new biological imputation benchmark, show that FADTI improves accuracy, uncertainty estimation, and sampling efficiency, especially under high missing rates and structured missing patterns. Code is available at https://github.com/RazeenLI/FADTI
Abstract: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. Existing Transformer- and diffusion-based imputers achieve strong performance, but they often rely mainly on time-domain modeling and lack adaptive spectral bias for recovering structured temporal gaps. We propose FADTI, a Fourier- and attention-driven diffusion framework for multivariate time series imputation. FADTI introduces a Fourier Bias Projection (FBP) module that injects learnable frequency-aware bias into intermediate hidden states during denoising. It projects intermediate hidden states onto Fourier bases, avoiding direct spectral estimation from masked or zero-filled inputs. With DFT, STFT, and FSST instantiations, FBP captures global periodicity, localized time--frequency variations, and non-stationary oscillatory patterns. By coupling FBP with self-attention and gated convolution, FADTI integrates frequency-domain guidance, temporal dependency modeling, and probabilistic denoising in a unified framework. Experiments on multiple benchmarks, including a new biological imputation benchmark, show that FADTI improves accuracy, uncertainty estimation, and sampling efficiency, especially under high missing rates and structured missing patterns. Code is available at this https URL
Accepted at the 2026 IEEE International Conference on Data Mining (ICDM 2026). 10 pages, 7 figures
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2512.15116 [cs.LG]
(or arXiv:2512.15116v3 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2512.15116 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 289 words
Related coverage
FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation
arXiv · Jun 9, 2026
SpikF-GO: Spiking Fourier Graph Operators for Multivariate Time Series Forecasting
arXiv · Jun 15, 2026
Knowledge-Assisted Multi-Graph Dependency Learning for Multivariate Time Series Anomaly Detection in Multi-Stage Industrial Processes
arXiv · Jul 20, 2026
[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
Latent Space · Sep 17, 2026