Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
This video presentation explores "AdaWorld," a groundbreaking AI foundation model that transcends traditional Large Language Models by learning directly from visual data. Discover how this innovative system extracts action understanding from unlabeled videos and transfers that knowledge across different environments. The 21-minute talk explains how AdaWorld creates adaptable AI that perceives and acts with remarkable efficiency, representing a paradigm shift in how intelligent agents comprehend their surroundings. Learn about continuous latent action learning and new world model approaches developed by researchers from HKUST, Harvard University, Google DeepMind, UMass Amherst, and the MIT-IBM Watson AI Lab. The presentation offers insights into the future of AI systems that can truly understand and interact with the physical world.
