World Models

Can agents learn inside of their own dreams?

--> --> This page requires Javascript. Please enable it for https://worldmodels.github.io/ Loading World Models ... Interactive demo: Tap screen to override the agent's decisions. scroll down --> An agent playing inside of a game environment hallucinated by a recurrent neural network. --> World Models Can agents learn inside of their own dreams? David Ha Google Brain Tokyo, Japan Jürgen Schmidhuber NNAISENSE Swiss AI Lab, IDSIA (USI & SUPSI) March 27 2018 NIPS 2018 Paper YouTube Talk Download PDF Abstract We explore building generative neural network models of popular reinforcement learning en

related reading

• World Models | Rohit Bandarurohitbandaru.github.io
• pdfopenreview.net
• World Models: Computing the Uncomputablenotboring.co
• [2509.24527] Training Agents Inside of Scalable World Modelsarxiv.org
• CIS 6280 · World Modelscis.upenn.edu
• Dream to Control: Learning Behaviors by Latent Imaginationdanijar.com
• Introducing Dreamer: Scalable Reinforcement Learning Using World Modelsresearch.google
• Efficient World Models with Context-Aware Tokenizationarxiv.org
• State of Robot Learning, December 2025vedder.io
• How not to do research - Rajan Agarwalrajan.sh
• A Functional Taxonomy of World Models - Dr. Fei-Fei Lidrfeifei.substack.com
• LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixelsle-wm.github.io