Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models
The paper titled 'Baba in Wonderland' explores online self-supervised dynamics discovery for executable world models. It introduces a system named Alice that refines candidate updates to improve learning under prior misalignment. The experiments demonstrate significant advancements in executable world-model learning through the proposed methods.
- ▪Executable world models can be edited and reused for planning if they accurately capture the environment's transition laws.
- ▪The system Alice treats failed candidate updates as signals to refine hypothesis classes and guide exploration.
- ▪Experiments on a variant of Baba Is You show that Alice enhances learning under conditions of prior misalignment.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.16725 (cs) [Submitted on 16 May 2026] Title:Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models Authors:SeungWon Seo, DongHeun Han, SeongRae Noh, HyeongYeop Kang View a PDF of the paper titled Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models, by SeungWon Seo and 3 other authors View PDF HTML (experimental) Abstract:Executable world models can be read, edited, executed, and reused for planning, but only if the program captures the environment's transition law rather than semantic shortcuts in its surface vocabulary.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
