Concept多模态信号输入
Multimodal input
A technical paradigm in which an AI model simultaneously receives multiple signal types, such as text, images, audio and video, and understands them jointly.
Timeline (last 90 days)
Oct 14
Multimodal input is not as simple as "stacking on features"; when the primary channel fails, the backup channel must be robust enough, otherwise it is merely failing in a different way.
Unverified50%