Deepmind Unisim Simulates Reality To Train Robots, Game Characters

Researchers at Google DeepMind, in collaboration with UC Berkeley, MIT, and the University of Alberta, have developed a new machine learning model called UniSim. This generative AI system can create realistic simulations for training all kinds of AI systems, and can be used to train embodied planners, low-level control, and other AI systems that require intricate real-world interactions. UniSim can simulate the visual outcomes of both high-level instructions and low-level controls, and can merge the vast knowledge contained in its training data to generalize beyond its training examples.