Align While Search: Belief-Guided Exploratory Inference for World-Grounded Embodied Agents
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914227552256000 |
|---|---|
| author | Bae, Seohui Kim, Jeonghye Sung, Youngchul Lim, Woohyung |
| author_facet | Bae, Seohui Kim, Jeonghye Sung, Youngchul Lim, Woohyung |
| contents | In this paper, we propose a test-time adaptive agent that performs exploratory inference through posterior-guided belief refinement without relying on gradient-based updates or additional training for LLM agent operating under partial observability. Our agent maintains an external structured belief over the environment state, iteratively updates it via action-conditioned observations, and selects actions by maximizing predicted information gain over the belief space. We estimate information gain using a lightweight LLM-based surrogate and assess world alignment through a novel reward that quantifies the consistency between posterior belief and ground-truth environment configuration. Experiments show that our method outperforms inference-time scaling baselines such as prompt-augmented or retrieval-enhanced LLMs, in aligning with latent world states with significantly lower integration overhead. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_24461 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Align While Search: Belief-Guided Exploratory Inference for World-Grounded Embodied Agents Bae, Seohui Kim, Jeonghye Sung, Youngchul Lim, Woohyung Artificial Intelligence In this paper, we propose a test-time adaptive agent that performs exploratory inference through posterior-guided belief refinement without relying on gradient-based updates or additional training for LLM agent operating under partial observability. Our agent maintains an external structured belief over the environment state, iteratively updates it via action-conditioned observations, and selects actions by maximizing predicted information gain over the belief space. We estimate information gain using a lightweight LLM-based surrogate and assess world alignment through a novel reward that quantifies the consistency between posterior belief and ground-truth environment configuration. Experiments show that our method outperforms inference-time scaling baselines such as prompt-augmented or retrieval-enhanced LLMs, in aligning with latent world states with significantly lower integration overhead. |
| title | Align While Search: Belief-Guided Exploratory Inference for World-Grounded Embodied Agents |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2512.24461 |