1
Fusion is not one-size-fits-all: Cross-Modal Representation Alignment for Time-to-Event Modeling
告别一刀切融合,提出针对时间事件建模的跨模态表示对齐新范式,精准提升多模态生存预测性能。
arXiv:2606.15038v1 Announce Type: new Abstract: Accurate time-to-event (TTE) prediction from multimodal clinical data remains challenging due to modal…