How can we teach AI skills using visual demos? Model-based Inverse Reinforcement Learning
发布时间 2021-01-27 15:25:37 来源
摘要
Most machines need hand-programmed rewards to pick up and place a bottle, for instance. Our new system learns a model of its environment, observes human behavior, and develops its own reward system. Learn more here: https://ai.facebook.com/blog/teaching-ai-to-manipulate-objects-using-visual-demos/
GPT-4正在为你翻译摘要中......
