Discovering Hybrid World Representations with Co-Evolving Foundation Models

Authors

  • Jiajun Wu Stanford University
  • Yunzhi Zhang Stanford University
  • Hong-Xing Yu Stanford University
  • Joy Hsu Stanford University
  • Jiayuan Mao University of Pennsylvania

DOI:

https://doi.org/10.1609/aaai.v40i48.42138

Abstract

This perspective article discusses an emerging research direction: to what extent can foundation models yield usable structure for modeling the physical world? We offer a Markovian formulation of structured world models and outline the notion of multi-level hybrid world representations that support compositional structure. We then review and suggest possible discovery paradigms, spanning distillation, interaction-driven continual learning, and ensemble learning.
AAAI-26 / IAAI-26 / EAAI-26 Proceedings Cover

Downloads

Published

2026-03-14

How to Cite

Wu, J., Zhang, Y., Yu, H.-X., Hsu, J., & Mao, J. (2026). Discovering Hybrid World Representations with Co-Evolving Foundation Models. Proceedings of the AAAI Conference on Artificial Intelligence, 40(48), 41020–41024. https://doi.org/10.1609/aaai.v40i48.42138

Issue

Vol. 40 No. 48: EAAI-26 AI for Education, Model AI Assignments, AAAI-26 Emerging Trends, Doctoral Consortium, Student Abstracts, Undergraduate Consortium and Demonstrations

Section

AAAI Emerging Trends in AI