Girish Dwivedi

Watch and Listen: Understanding Audio-Visual-Speech Moments with Multimodal LLM

Humans naturally understand moments in a video by integrating visual and auditory cues. For example, localizing a scene in the video like “A scientist passionately speaks on wildlife conservation as dramatic orchestral music plays, with the audience nodding and applauding” requires simultaneous proc…