ToolACE-DEV: Self-Improving Tool Learning via Decomposition and EVolution
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_ | 1866915283701071872 |
|---|---|
| author | Huang, Xu Liu, Weiwen Zeng, Xingshan Huang, Yuefeng Hao, Xinlong Wang, Yuxian Zeng, Yirong Wu, Chuhan Wang, Yasheng Tang, Ruiming Lian, Defu |
| author_facet | Huang, Xu Liu, Weiwen Zeng, Xingshan Huang, Yuefeng Hao, Xinlong Wang, Yuxian Zeng, Yirong Wu, Chuhan Wang, Yasheng Tang, Ruiming Lian, Defu |
| contents | The tool-using capability of large language models (LLMs) enables them to access up-to-date external information and handle complex tasks. Current approaches to enhancing this capability primarily rely on distilling advanced models by data synthesis. However, this method incurs significant costs associated with advanced model usage and often results in data compatibility issues, led by the high discrepancy in the knowledge scope between the advanced model and the target model. To address these challenges, we propose ToolACE-DEV, a self-improving framework for tool learning. First, we decompose the tool-learning objective into sub-tasks that enhance basic tool-making and tool-using abilities. Then, we introduce a self-evolving paradigm that allows lightweight models to self-improve, reducing reliance on advanced LLMs. Extensive experiments validate the effectiveness of our approach across models of varying scales and architectures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_07512 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ToolACE-DEV: Self-Improving Tool Learning via Decomposition and EVolution Huang, Xu Liu, Weiwen Zeng, Xingshan Huang, Yuefeng Hao, Xinlong Wang, Yuxian Zeng, Yirong Wu, Chuhan Wang, Yasheng Tang, Ruiming Lian, Defu Computation and Language Artificial Intelligence The tool-using capability of large language models (LLMs) enables them to access up-to-date external information and handle complex tasks. Current approaches to enhancing this capability primarily rely on distilling advanced models by data synthesis. However, this method incurs significant costs associated with advanced model usage and often results in data compatibility issues, led by the high discrepancy in the knowledge scope between the advanced model and the target model. To address these challenges, we propose ToolACE-DEV, a self-improving framework for tool learning. First, we decompose the tool-learning objective into sub-tasks that enhance basic tool-making and tool-using abilities. Then, we introduce a self-evolving paradigm that allows lightweight models to self-improve, reducing reliance on advanced LLMs. Extensive experiments validate the effectiveness of our approach across models of varying scales and architectures. |
| title | ToolACE-DEV: Self-Improving Tool Learning via Decomposition and EVolution |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2505.07512 |