EvoEmpirBench: Dynamic Spatial Reasoning with Agent-ExpVer
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909925189353472 |
|---|---|
| author | Zhao, Pukun Wang, Longxiang Wang, Miaowei Chen, Chen Zhou, Fanqing Huang, Haojian |
| author_facet | Zhao, Pukun Wang, Longxiang Wang, Miaowei Chen, Chen Zhou, Fanqing Huang, Haojian |
| contents | Most existing spatial reasoning benchmarks focus on static or globally observable environments, failing to capture the challenges of long-horizon reasoning and memory utilization under partial observability and dynamic changes. We introduce two dynamic spatial benchmarks, locally observable maze navigation and match-2 elimination that systematically evaluate models' abilities in spatial understanding and adaptive planning when local perception, environment feedback, and global objectives are tightly coupled. Each action triggers structural changes in the environment, requiring continuous update of cognition and strategy. We further propose a subjective experience-based memory mechanism for cross-task experience transfer and validation. Experiments show that our benchmarks reveal key limitations of mainstream models in dynamic spatial reasoning and long-term memory, providing a comprehensive platform for future methodological advances. Our code and data are available at https://anonymous.4open.science/r/EvoEmpirBench-143C/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_12718 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | EvoEmpirBench: Dynamic Spatial Reasoning with Agent-ExpVer Zhao, Pukun Wang, Longxiang Wang, Miaowei Chen, Chen Zhou, Fanqing Huang, Haojian Computer Vision and Pattern Recognition Most existing spatial reasoning benchmarks focus on static or globally observable environments, failing to capture the challenges of long-horizon reasoning and memory utilization under partial observability and dynamic changes. We introduce two dynamic spatial benchmarks, locally observable maze navigation and match-2 elimination that systematically evaluate models' abilities in spatial understanding and adaptive planning when local perception, environment feedback, and global objectives are tightly coupled. Each action triggers structural changes in the environment, requiring continuous update of cognition and strategy. We further propose a subjective experience-based memory mechanism for cross-task experience transfer and validation. Experiments show that our benchmarks reveal key limitations of mainstream models in dynamic spatial reasoning and long-term memory, providing a comprehensive platform for future methodological advances. Our code and data are available at https://anonymous.4open.science/r/EvoEmpirBench-143C/. |
| title | EvoEmpirBench: Dynamic Spatial Reasoning with Agent-ExpVer |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.12718 |