Get the latest tech news
'World Models,' an old idea in AI, mount a comeback
You’re carrying around in your head a model of how the world works. Will AI systems need to do the same?
The deep learning luminaries Yann LeCun (of Meta), Demis Hassabis (of Google DeepMind) and Yoshua Bengio (of Mila, the Quebec Artificial Intelligence Institute) all believe world models are essential for building AI systems that are truly smart, scientific and safe. Instead of relying on brittle hand-coded rules, deep neural networks could build up internal approximations of their training environments through trial and error and then use them to accomplish narrowly specified tasks, such as driving a virtual race car. To prominent AI experts such as Geoffrey Hinton, Ilya Sutskever and Chris Olah, it was obvious: Buried somewhere deep within an LLM’s thicket of virtual neurons must lie “a small-scale model of external reality,” just as Craik imagined.
Or read this on Hacker News