Aligning Agentic World Models via Knowledgeable Experience Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912833492484096 |
|---|---|
| author | Ren, Baochang Yao, Yunzhi Sun, Rui Qiao, Shuofei Zhang, Ningyu Chen, Huajun |
| author_facet | Ren, Baochang Yao, Yunzhi Sun, Rui Qiao, Shuofei Zhang, Ningyu Chen, Huajun |
| contents | Current Large Language Models (LLMs) exhibit a critical modal disconnect: they possess vast semantic knowledge but lack the procedural grounding to respect the immutable laws of the physical world. Consequently, while these agents implicitly function as world models, their simulations often suffer from physical hallucinations-generating plans that are logically sound but physically unexecutable. Existing alignment strategies predominantly rely on resource-intensive training or fine-tuning, which attempt to compress dynamic environmental rules into static model parameters. However, such parametric encapsulation is inherently rigid, struggling to adapt to the open-ended variability of physical dynamics without continuous, costly retraining. To bridge this gap, we introduce WorldMind, a framework that autonomously constructs a symbolic World Knowledge Repository by synthesizing environmental feedback. Specifically, it unifies Process Experience to enforce physical feasibility via prediction errors and Goal Experience to guide task optimality through successful trajectories. Experiments on EB-ALFRED and EB-Habitat demonstrate that WorldMind achieves superior performance compared to baselines with remarkable cross-model and cross-environment transferability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_13247 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Aligning Agentic World Models via Knowledgeable Experience Learning Ren, Baochang Yao, Yunzhi Sun, Rui Qiao, Shuofei Zhang, Ningyu Chen, Huajun Computation and Language Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning Multimedia Current Large Language Models (LLMs) exhibit a critical modal disconnect: they possess vast semantic knowledge but lack the procedural grounding to respect the immutable laws of the physical world. Consequently, while these agents implicitly function as world models, their simulations often suffer from physical hallucinations-generating plans that are logically sound but physically unexecutable. Existing alignment strategies predominantly rely on resource-intensive training or fine-tuning, which attempt to compress dynamic environmental rules into static model parameters. However, such parametric encapsulation is inherently rigid, struggling to adapt to the open-ended variability of physical dynamics without continuous, costly retraining. To bridge this gap, we introduce WorldMind, a framework that autonomously constructs a symbolic World Knowledge Repository by synthesizing environmental feedback. Specifically, it unifies Process Experience to enforce physical feasibility via prediction errors and Goal Experience to guide task optimality through successful trajectories. Experiments on EB-ALFRED and EB-Habitat demonstrate that WorldMind achieves superior performance compared to baselines with remarkable cross-model and cross-environment transferability. |
| title | Aligning Agentic World Models via Knowledgeable Experience Learning |
| topic | Computation and Language Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning Multimedia |
| url | https://arxiv.org/abs/2601.13247 |