T1: One-to-One Channel-Head Binding for Multivariate Time-Series Imputation

2026-02-24

AI Summary

This work addresses the challenge of imputing multivariate time series under high missing rates and complex missing patterns, where existing methods suffer from degraded temporal features that hinder effective cross-variable information transfer and lead to large reconstruction errors. To overcome these limitations, the authors propose the T1 model, which employs a CNN-Transformer hybrid architecture featuring an innovative one-to-one binding mechanism between channels and attention heads. This design jointly optimizes temporal feature extraction and cross-variable dependency modeling, while adaptive attention weighting suppresses pathways corrupted by missing data. Evaluated on 11 benchmark datasets, T1 achieves state-of-the-art performance, reducing average MSE by 46% compared to the next-best method, maintains robustness even at 70% missingness, and demonstrates zero-shot generalization to unseen missing patterns without retraining.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Imputing missing values in multivariate time series remains challenging, especially under diverse missing patterns and heavy missingness. Existing methods suffer from suboptimal performance as corrupted temporal features hinder effective cross-variable information transfer, amplifying reconstruction errors. Robust imputation requires both extracting temporal patterns from sparse observations within each variable and selectively transferring information across variables--yet current approaches excel at one while compromising the other. We introduce T1 (Time series imputation with 1-to-1 channel-head binding), a CNN-Transformer hybrid architecture that achieves robust imputation through Channel-Head Binding--a mechanism creating one-to-one correspondence between CNN channels and attention heads. This design enables selective information transfer: when missingness corrupts certain temporal patterns, their corresponding attention pathways adaptively down-weight based on remaining observable patterns while preserving reliable cross-variable connections through unaffected channels. Experiments on 11 benchmark datasets demonstrate that T1 achieves state-of-the-art performance, reducing MSE by 46% on average compared to the second-best baseline, with particularly strong gains under extreme sparsity (70% missing ratio). The model generalizes to unseen missing patterns without retraining and uses a consistent hyperparameter configuration across all datasets. The code is available at https://github.com/Oppenheimerdinger/T1.

Problem

• multivariate time-series imputation
• missing data
• temporal patterns
• cross-variable information transfer
• missingness

Innovation

• Channel-Head Binding
• multivariate time-series imputation
• CNN-Transformer hybrid
• selective cross-variable transfer
• robust imputation under sparsity