Tag
This paper challenges the conventional repair-first paradigm for missing modalities in multimodal sentiment analysis, showing that full-modality input is only optimal for a small fraction of samples. The authors propose SIEVE, a plug-and-play method that learns sample-level decisions on whether to repair missing modalities, consistently improving existing repair backbones.
RL4IL introduces a reinforcement learning-guided retrieval method that uses soft fusion over frozen demonstration libraries to handle missing sensor modalities in robotic imitation learning at inference time, achieving high success rates under complete camera dropout.
LongMoE proposes a unified framework that jointly addresses modality missingness and longitudinal dynamics in multimodal clinical learning, using context-aware imputation, attentional tokenization, trajectory-aware encoding, and sparse mixture-of-experts routing. Experiments on ADNI, OASIS-3, and MIMIC-IV demonstrate improved robustness under missing modalities while remaining competitive in full-modality settings.
This paper proposes CL-DMDF, a dynamic multimodal data fusion model that uses contrastive learning and a dual-dimensional attention mechanism to handle missing modalities and improve discriminative learning.
This paper proposes a graph-based one-stage framework for brain tumor segmentation that handles missing MRI modalities by introducing modality-specific virtual nodes and a dynamic connection strategy, outperforming state-of-the-art methods on the BRATS-2018 and BRATS-2020 datasets.
MuteBench is a benchmark for evaluating multimodal fusion models under modality missing and within-modality missing conditions across clinical datasets. It provides insights into architecture robustness and suggests that diffusion-based imputation can help.
This paper proposes FedMPO, a robust federated multimodal graph learning method that addresses modality heterogeneity and missing modalities through topology-aware cross-modal generation, missing-aware expert routing, and reliability-aware aggregation, achieving performance gains on multiple datasets.