Researchers from NVIDIA, in collaboration with MIT and the University of Oxford, have introduced Physis-Lang, a self-evolving physical language designed to enhance the realism of physics in video world models. This innovative approach aims to address common issues where visual representations often violate physical laws, such as objects behaving unrealistically.
Physis-Lang treats physical language as a shared and optimizable framework, focusing on leveraging language rather than relying solely on visual or numerical signals. This advancement is expected to significantly improve the accuracy of physics simulations in video content, offering a more immersive experience for users.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
