1 paper
Xinxin Chen, Yuchen Li, Zihan Wang +3
Current multimodal fusion approaches, particularly those based on static Mixture-of-Experts (MoE) architectures, often struggle to provide the adaptive and efficient collaborative…