Represent, Then Generate: Multimodal-Conditioned Time-Series Generation under Irregular Missingness

2026-08-12

AI Summary

This study addresses the challenge of missing critical physiological signals in clinical monitoring—often due to invasiveness, high cost, or unavailability—and the limitations of existing methods in handling multimodal, irregularly missing time-series data alongside static covariates. The authors propose ReCoGen, a two-stage framework: first, a masked autoencoder extracts compact, missingness-robust token sequences for each modality; second, a flow-matching generator synthesizes target signals by fusing these tokens with static conditions. The approach innovatively decouples conditional representation from generation and introduces a learnable cross-attention mechanism along with a dual-path fusion strategy for static conditions (via tokens and AdaLN). Evaluated across 16 tasks on AI-READI, MIMIC-III, and MIMIC-IV, ReCoGen achieves state-of-the-art downstream utility, surpassing even real-signal baselines in 13 cases.

Abstract

Continuous physiological time series underpin modern clinical monitoring, yet many of the most informative signals are invasive, expensive, or simply unavailable for a given patient. Conditional generation offers a remedy: an absent signal can be synthesized from co-recorded signals and routine clinical variables. Existing generators, however, are built around a single conditioning modality and degrade when forced to handle the heterogeneous, irregularly missing mix of time-variant signals and static covariates seen in practice. We propose ReCoGen (Represent Conditions, then Generate), a two-stage framework that decouples multimodal condition representation from target generation. Stage I trains one masked autoencoder per modality, distilling each time-variant condition into a compact and missingness-tolerant token sequence. Stage II trains a flow-matching generator that fuses these tokens with static conditions to synthesize the target signal. Across three physiological benchmarks, including continuous glucose monitoring on AI-READI and arterial blood pressure generation on MIMIC-III and MIMIC-IV, ReCoGen attains the best downstream utility on all sixteen (dataset, task, metric) settings, surpassing six representative conditional generators; on thirteen of them its utility also reaches or exceeds the utility measured on the real signal, a reference we read as an approximate anchor rather than a ceiling. Ablations trace the gains to the conditioning path: learnable cross-attention over the frozen per-modality encoders, and a dual token-plus-AdaLN route for the static conditions. ReCoGen thus turns routinely collected signals into informative surrogates for invasive or unavailable ones, a step toward less invasive, lower-cost continuous clinical monitoring.

Problem

• time-series generation
• irregular missingness
• multimodal conditioning
• physiological signals
• clinical monitoring

Innovation

• multimodal-conditioned generation
• irregular missingness
• masked autoencoder
• flow-matching
• two-stage framework