Toward Communication-Efficient and Data-Free Collaborative Fine-Tuning Between Small and Large Language Models
IEEE Transactions on Networking · 2026 · 34 · 5333-5348
DOI: 10.1109/ton.2026.3695944 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Networking, 2026. Verfasst von Zhenya Ma und 5 weiteren Autor:innen. Die Arbeit nennt 68 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China, China Postdoctoral Science Foundation.
Abstract (Originalsprache)
While large language models (LLMs) exhibit impressive general capabilities, their performance on domain-specific tasks often requires fine-tuning with private data that cannot be shared due to privacy constraints. Directly deploying LLMs on resource-constrained clients for local fine-tuning is impractical due to their significant computation and communication costs. In addition, pre-trained LLMs are valuable intellectual property, and model owners are reluctant to distribute full model weights. To address these challenges, we propose CoT-LM, a communication-efficient, computation-light, and data-free framework for collaborative fine-tuning between small (SLMs) and large language models (LLMs). In CoT-LM, clients fine-tune lightweight SLMs locally without uploading models or private data. These SLMs provide task-specific feedback to guide server-side LLM enhancement via an efficient communication protocol that exchanges only lightweight synthetic data and feedback. The framework supports both synchronous and asynchronous collaboration and enables mutual enhancement: the LLM improves its task-specific capabilities, while clients benefit from refined synthetic data or distilled knowledge. Extensive experiments demonstrate that CoT-LM significantly boosts natural language understanding (NLU, up to 18.3% for LLMs and 8.0% for SLMs) and natural language generation (NLG, up to 31.7% for LLMs) performance across diverse tasks while preserving data privacy, model intellectual property, and generalization capabilities, achieving significant reductions in computation and communication overhead.
Förderung
- National Natural Science Foundation of China
- Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China
- China Postdoctoral Science Foundation
- Natural Science Foundation of Guangdong Province
- Shenzhen Science and Technology Program
- Young Elite Scientist Sponsorship Program by China Association for Science and Technology
- Guoqiang Institute, Tsinghua University
Zitieren
APA
Ma, Z., Deng, Y., Qiao, Z., Fu, Y., Yue, S., & Ren, J. (2026). Toward Communication-Efficient and Data-Free Collaborative Fine-Tuning Between Small and Large Language Models. IEEE Transactions on Networking, 34, 5333-5348. https://doi.org/10.1109/ton.2026.3695944
Vancouver
Ma Z, Deng Y, Qiao Z, Fu Y, Yue S, Ren J. Toward Communication-Efficient and Data-Free Collaborative Fine-Tuning Between Small and Large Language Models. IEEE Transactions on Networking. 2026;34:5333-5348. doi:10.1109/ton.2026.3695944
BibTeX
@article{Ma2026,
author = {Ma, Zhenya and Deng, Yongheng and Qiao, Ziqing and Fu, Yongjian and Yue, Sheng and Ren, Ju},
title = {Toward Communication-Efficient and Data-Free Collaborative Fine-Tuning Between Small and Large Language Models},
journal = {IEEE Transactions on Networking},
year = {2026},
volume = {34},
pages = {5333--5348},
doi = {10.1109/ton.2026.3695944}
}
EUVIMED