Sanjoy Chowdhury

A method for robots to generate wayfinding instructions using language models.

New model generates music using both text and visual information.

A new model combines audio and visual data for improved understanding.