AB  - The Internet of Things technology has evolved from standalone tools to open systems supporting parallel and distributed computing. In particular, federated learning (FL) has become a key solution as a distributed team of parallel training methods in the artificial intelligence of things (AIoT). Academia and industry formed a significant consensus on efficient reproducing and rapidly deploying federated learning in the AIoT. The current best practice typically resorts to three approaches: 1) looking for publicly open-source algorithmic prototypes, 2) contacting the authors to get a private prototype, and 3) manually implementing a prototype following the description of the publication. However, most published network research does not have public prototypes, and private prototypes are hard to get. As such, most reproducing efforts are spent on manual implementation based on the publications, which is both time and labor-consuming and error-prone. In this paper, we propose reproducing FL research results using the emerging large language models (LLMs). In particular, we first prove its feasibility with an experiment in which three students with essential parallel and distributed machine learning knowledge reproduce different FL algorithms published in prominent conferences and journals by prompt engineering ChatGPT-4. Finally, our experimental results focus on the efficiency and quality of reproducing code. We report the experiment?s observations and discuss future open research questions of this paper. Additionally, we verify the better robustness of reproduced codes with different data poisoning attacks via extensive experiments. This work also raises no ethical issue.
AU  - Du, H
AU  - Li, W
AU  - Ding, X
AU  - Huo, H
DA  - 2024/11/10
DO  - 10.1109/acait63902.2024.11022196
EP  - 510
JO  - 2024 8th Asian Conference on Artificial Intelligence Technology (ACAIT)
PB  - Institute of Electrical and Electronics Engineers (IEEE)
PY  - 2024/11/10
SP  - 501
TI  - Can LLMs Understand Parallel and Distributed Machine Learning Algorithms in AIoT?
VL  - 00
Y1  - 2024/11/10
Y2  - 2026/05/15
ER  -