Event-Centric World Modeling with Memory-Augmented Retrieval for Embodied Decision-Making

Zhaowen Fan, Rongchao Zhang · 2026

Read Event-Centric World Modeling with Memory-Augmented Retrieval for Embodied Decision-Making by Zhaowen Fan, Rongchao Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07392.txt.

Description

Fan, Zhang, and Han propose ERA, a clear framework for embodied choice that beats black-box imitation. The world is stored as structured semantic events in a readable form. Choices come from pulling similar past event-action pairs from a store of cases. Final moves blend those retrieved moves by weight, keeping choices physical and open to check. Drone tests show real-time, adaptive moves in changing settings with no retraining.

Catalog record

CREATORZhaowen Fan, Rongchao Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41153
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.07392

Browse the full library · About the Tower of Babel