Robust Multimodal Representation Learning in Healthcare
This work addresses biases in healthcare data integration for clinical outcome prediction, representing an incremental improvement over existing methods that neglect biased features.
The paper tackles the problem of systematic biases in medical multimodal representation learning by proposing a Dual-Stream Feature Decorrelation Framework, which disentangles causal features from spurious correlations, resulting in consistent performance improvements across MIMIC-IV, eICU, and ADNI datasets.
Medical multimodal representation learning aims to integrate heterogeneous data into unified patient representations to support clinical outcome prediction. However, real-world medical datasets commonly contain systematic biases from multiple sources, which poses significant challenges for medical multimodal representation learning. Existing approaches typically focus on effective multimodal fusion, neglecting inherent biased features that affect the generalization ability. To address these challenges, we propose a Dual-Stream Feature Decorrelation Framework that identifies and handles the biases through structural causal analysis introduced by latent confounders. Our method employs a causal-biased decorrelation framework with dual-stream neural networks to disentangle causal features from spurious correlations, utilizing generalized cross-entropy loss and mutual information minimization for effective decorrelation. The framework is model-agnostic and can be integrated into existing medical multimodal learning methods. Comprehensive experiments on MIMIC-IV, eICU, and ADNI datasets demonstrate consistent performance improvements.