Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.21902 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914114130935808 |
|---|---|
| author | Boulet, Timothé Hinaut, Xavier Moulin-Frier, Clément |
| author_facet | Boulet, Timothé Hinaut, Xavier Moulin-Frier, Clément |
| contents | Software Engineering Agents (SWE-Agents) have proven effective for traditional software engineering tasks with accessible codebases, but their performance for embodied tasks requiring well-designed information discovery remains unexplored. We present the first extended evaluation of SWE-Agents on controller generation for embodied tasks, adapting Mini-SWE-Agent (MSWEA) to solve 20 diverse embodied tasks from the Minigrid environment. Our experiments compare agent performance across different information access conditions: with and without environment source code access, and with varying capabilities for interactive exploration. We quantify how different information access levels affect SWE-Agent performance for embodied tasks and analyze the relative importance of static code analysis versus dynamic exploration for task solving. This work establishes controller generation for embodied tasks as a crucial evaluation domain for SWE-Agents and provides baseline results for future research in efficient reasoning systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_21902 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Software Engineering Agents for Embodied Controller Generation : A Study in Minigrid Environments Boulet, Timothé Hinaut, Xavier Moulin-Frier, Clément Software Engineering Artificial Intelligence Software Engineering Agents (SWE-Agents) have proven effective for traditional software engineering tasks with accessible codebases, but their performance for embodied tasks requiring well-designed information discovery remains unexplored. We present the first extended evaluation of SWE-Agents on controller generation for embodied tasks, adapting Mini-SWE-Agent (MSWEA) to solve 20 diverse embodied tasks from the Minigrid environment. Our experiments compare agent performance across different information access conditions: with and without environment source code access, and with varying capabilities for interactive exploration. We quantify how different information access levels affect SWE-Agent performance for embodied tasks and analyze the relative importance of static code analysis versus dynamic exploration for task solving. This work establishes controller generation for embodied tasks as a crucial evaluation domain for SWE-Agents and provides baseline results for future research in efficient reasoning systems. |
| title | Software Engineering Agents for Embodied Controller Generation : A Study in Minigrid Environments |
| topic | Software Engineering Artificial Intelligence |
| url | https://arxiv.org/abs/2510.21902 |