Abstract: Multimodal intent recognition utilizes heterogeneous modalities such as visual, auditory, and textual cues to infer user intent, serving as a pivotal component in human-machine interaction.

Abstract: Data-driven prognostics and health management techniques are increasingly applied by performance prediction, but the performance of deep learning models is limited by physical data scarcity.