Visual Attention Reasoning via Hierarchical Search and Self-Verification
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914279446282240 |
|---|---|
| author | Cai, Wei Zhao, Jian Yuan, Yuchen Zhang, Tianle Zhu, Ming Tang, Haichuan Li, Xuelong |
| author_facet | Cai, Wei Zhao, Jian Yuan, Yuchen Zhang, Tianle Zhu, Ming Tang, Haichuan Li, Xuelong |
| contents | Multimodal Large Language Models (MLLMs) frequently hallucinate due to their reliance on fragile, linear reasoning and weak visual grounding. We propose Visual Attention Reasoning (VAR), a reinforcement learning framework that reformulates reasoning as a hierarchical search with self-verification. VAR enforces traceable evidence grounding by generating explicit bounding boxes, guided by a novel reward function combining geometric precision and semantic sufficiency. Furthermore, it replaces linear Chain-of-Thought with a tree-search policy capable of backtracking to correct logical errors. Theoretical analysis validates the framework's reliability, and extensive experiments demonstrate that VAR significantly outperforms state-of-the-art methods on complex hallucination and safety benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_18619 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Visual Attention Reasoning via Hierarchical Search and Self-Verification Cai, Wei Zhao, Jian Yuan, Yuchen Zhang, Tianle Zhu, Ming Tang, Haichuan Li, Xuelong Artificial Intelligence Multimodal Large Language Models (MLLMs) frequently hallucinate due to their reliance on fragile, linear reasoning and weak visual grounding. We propose Visual Attention Reasoning (VAR), a reinforcement learning framework that reformulates reasoning as a hierarchical search with self-verification. VAR enforces traceable evidence grounding by generating explicit bounding boxes, guided by a novel reward function combining geometric precision and semantic sufficiency. Furthermore, it replaces linear Chain-of-Thought with a tree-search policy capable of backtracking to correct logical errors. Theoretical analysis validates the framework's reliability, and extensive experiments demonstrate that VAR significantly outperforms state-of-the-art methods on complex hallucination and safety benchmarks. |
| title | Visual Attention Reasoning via Hierarchical Search and Self-Verification |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2510.18619 |