MissHDD: Hybrid Deterministic Diffusion for Hetrogeneous Incomplete Data Imputation

2025-11-18

AI Summary

Real-world tabular data often contain heterogeneous features—both numerical and categorical—whereas existing diffusion models assume homogeneous feature spaces and rely on stochastic denoising, leading to poor conditional consistency, categorical mode collapse, and unstable numerical updates. To address this, we propose the Hybrid Deterministic Diffusion Framework (HDDF), introducing a novel dual-channel generation mechanism: numerical variables are modeled via conditional DDIM for deterministic denoising, while categorical variables leverage a discrete latent-path diffusion inspired by loopy belief propagation, jointly optimized under a unified conditional generation objective. This design explicitly decouples heterogeneous feature dynamics, balancing imputation accuracy and sampling stability. Experiments across multiple real-world datasets under MCAR, MAR, and MNAR missingness mechanisms demonstrate that HDDF significantly outperforms state-of-the-art diffusion-based and conventional imputation methods in accuracy, robustness, and convergence reliability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Feature Construction/Reformulation
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Incomplete data are common in real-world tabular applications, where numerical, categorical, and discrete attributes coexist within a single dataset. This heterogeneous structure presents significant challenges for existing diffusion-based imputation models, which typically assume a homogeneous feature space and rely on stochastic denoising trajectories. Such assumptions make it difficult to maintain conditional consistency, and they often lead to information collapse for categorical variables or instability when numerical variables require deterministic updates. These limitations indicate that a single diffusion process is insufficient for mixed-type tabular imputation. We propose a hybrid deterministic diffusion framework that separates heterogeneous features into two complementary generative channels. A continuous DDIM-based channel provides efficient and stable deterministic denoising for numerical variables, while a discrete latent-path diffusion channel, inspired by loopholing-based discrete diffusion, models categorical and discrete features without leaving their valid sample manifolds. The two channels are trained under a unified conditional imputation objective, enabling coherent reconstruction of mixed-type incomplete data. Extensive experiments on multiple real-world datasets show that the proposed framework achieves higher imputation accuracy, more stable sampling trajectories, and improved robustness across MCAR, MAR, and MNAR settings compared with existing diffusion-based and classical methods. These results demonstrate the importance of structure-aware diffusion processes for advancing deep learning approaches to incomplete tabular data.

Problem

• Imputing mixed-type tabular data with numerical and categorical variables
• Addressing instability in diffusion models for heterogeneous incomplete datasets
• Ensuring conditional consistency and avoiding information collapse in imputation

Innovation

• Hybrid deterministic diffusion separates heterogeneous features into two channels
• Continuous DDIM channel provides stable denoising for numerical variables
• Discrete latent-path diffusion models categorical features without leaving manifolds