Affordances Enable Partial World Modeling with LLMs
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918331958689792 |
|---|---|
| author | Khetarpal, Khimya Comanici, Gheorghe Richens, Jonathan Shar, Jeremy Xia, Fei Orseau, Laurent Faust, Aleksandra Precup, Doina |
| author_facet | Khetarpal, Khimya Comanici, Gheorghe Richens, Jonathan Shar, Jeremy Xia, Fei Orseau, Laurent Faust, Aleksandra Precup, Doina |
| contents | Full models of the world require complex knowledge of immense detail. While pre-trained large models have been hypothesized to contain similar knowledge due to extensive pre-training on vast amounts of internet scale data, using them directly in a search procedure is inefficient and inaccurate. Conversely, partial models focus on making high quality predictions for a subset of state and actions: those linked through affordances that achieve user intents~\citep{khetarpal2020can}. Can we posit large models as partial world models? We provide a formal answer to this question, proving that agents achieving task-agnostic, language-conditioned intents necessarily possess predictive partial-world models informed by affordances. In the multi-task setting, we introduce distribution-robust affordances and show that partial models can be extracted to significantly improve search efficiency. Empirical evaluations in tabletop robotics tasks demonstrate that our affordance-aware partial models reduce the search branching factor and achieve higher rewards compared to full world models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_10390 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Affordances Enable Partial World Modeling with LLMs Khetarpal, Khimya Comanici, Gheorghe Richens, Jonathan Shar, Jeremy Xia, Fei Orseau, Laurent Faust, Aleksandra Precup, Doina Machine Learning Artificial Intelligence Full models of the world require complex knowledge of immense detail. While pre-trained large models have been hypothesized to contain similar knowledge due to extensive pre-training on vast amounts of internet scale data, using them directly in a search procedure is inefficient and inaccurate. Conversely, partial models focus on making high quality predictions for a subset of state and actions: those linked through affordances that achieve user intents~\citep{khetarpal2020can}. Can we posit large models as partial world models? We provide a formal answer to this question, proving that agents achieving task-agnostic, language-conditioned intents necessarily possess predictive partial-world models informed by affordances. In the multi-task setting, we introduce distribution-robust affordances and show that partial models can be extracted to significantly improve search efficiency. Empirical evaluations in tabletop robotics tasks demonstrate that our affordance-aware partial models reduce the search branching factor and achieve higher rewards compared to full world models. |
| title | Affordances Enable Partial World Modeling with LLMs |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2602.10390 |