Residual diffusion model improves missing data recovery in time series
RDDMPI: Residual Denoising Diffusion Model for Probabilistic Multivariate Time Series Imputation
Machine Learning
Summary
Multivariate time series data often have missing values, which can be important to fill in accurately, such as in healthcare or traffic monitoring. The authors propose a method called RDDMPI that focuses on correcting just the uncertain parts left after an initial guess instead of trying to generate the entire missing data from scratch. This approach simplifies the problem and helps produce more accurate and reliable reconstructions. Their experiments show better results compared to previous methods.
What this means in practice
- •For healthcare data teams: Improve the accuracy and reliability of patient monitoring data by filling in missing sensor readings using residual diffusion imputation.
- •For energy systems analysts: Enhance energy usage datasets by probabilistically imputing missing values with better uncertainty estimates for more robust demand forecasting.
Authors
Ramiro Valdes Jara, David Chapman, Adam Meyers
Abstract
Multivariate time series imputation (MTSI) aims to recover missing values in temporal data composed of multiple interdependent variables. This problem is central to real-world applications such as healthcare monitoring, traffic networks, and energy systems. Recent diffusion-based approaches have shown strong potential for probabilistic imputation by learning to generate missing values through iterative denoising. However, most existing approaches perform diffusion directly in the original data space, requiring the denoising network to simultaneously capture global structure, temporal dynamics, and stochastic variability. This makes the generative task unnecessarily complex, especially when modern deterministic imputers can already provide accurate initial reconstructions. To address this limitation, we propose RDDMPI, a conditional residual diffusion framework that operates directly in residual space. Instead of modeling the full missing signal directly, we reformulate probabilistic imputation as a baseline-residual decomposition, where a pretrained model captures the dominant signal and a diffusion process models the residual uncertainty. To better exploit deterministic guidance, \model{} conditions the reverse denoising process on both the baseline-completed signal and its latent representation, while a reliability-aware conditioning mechanism adaptively controls the influence of baseline information during residual generation. This formulation simplifies the diffusion learning objective, enabling it to focus on structured correction terms rather than reconstructing the full signal. Experiments on multiple benchmark datasets demonstrate that RDDMPI consistently improves both reconstruction accuracy and uncertainty quantification.