The Case For World Models, Part I: The Neuroscientific Reason
Fei-Fei Li, Yann LeCun, Demis Hassabis and others are pushing for AIs that learn world models, to plan & predict accurately. Neuroscientists have known for decades that our brains must be doing this
credit: Substack image generator LOOK at the two images above. What do you see? In the left panel is a crater in the sand, with someone standing near the bottom. What about the right panel? Most of us are going to see something odd: the crater has turned into a mound, and the upright figure seems be lying flat on the sand. But here's the odder thing. The two images are exactly the same. The right image is simply the left one turned upside down. Nothing else has changed in terms of the raw information entering your eyes. (If you don't believe it, disable the autorotate feature on your…
saved by
related reading
- Trading places: What happens when neuroscience turns into machine learning, and machine learning turns into neuroscience?thetransmitter.org
- A Functional Taxonomy of World Models - Dr. Fei-Fei Lidrfeifei.substack.com
- Verbalizable Representations Form a Global Workspace in Language Modelstransformer-circuits.pub
- The World Inside Neural Networksgoodfire.ai
- Toy Models of Superpositiontransformer-circuits.pub
- Language Models, World Models, and Human Model-Buildinglingo.csail.mit.edu
- World Models | Rohit Bandarurohitbandaru.github.io
- LLMs and World Models, Part 1 - by Melanie Mitchellaiguide.substack.com
- World Models: Computing the Uncomputablenotboring.co
- Yann LeCun: World Models: Enabling the next AI revolutionyoutube.com
- What Is ChatGPT Doing … and Why Does It Work?-Stephen Wolfram Writingswritings.stephenwolfram.com
- Scaling works. These researchers are betting billions it isn't enoughsubstack.com