Exploring and Unleashing the Power of Large Language Models in Automated Code Translation
Proceedings of the ACM on Software Engineering · 12.07.2024 · 1 (FSE) · 1585-1608
DOI: 10.1145/3660778 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the ACM on Software Engineering, online erschienen am 12.07.2024. Verfasst von Zhen Yang und 9 weiteren Autor:innen. Die Arbeit wurde bisher 130-mal zitiert (Crossref) und nennt 67 Literaturangaben. Gefördert von: National Key R&D Program, National Natural Science Foundation of China, Shandong Province Overseas Outstanding Youth Fund.
Abstract (Originalsprache)
Code translation tools, namely transpilers, are developed for automatic source-to-source translation. Latest learning-based transpilers have shown impressive enhancement against rule-based counterparts in both translation accuracy and readability, owing to their task-specific pre-training on extensive monolingual corpora. Nevertheless, their current performance still remains unsatisfactory for practical deployment, and the associated training resources are also prohibitively expensive. Large Language Models (LLMs), pre-trained on huge amounts of human-written code/text, have shown remarkable performance in many code intelligence tasks due to their powerful generality, even without task-specific re-training/fine-tuning. Thus, LLMs can potentially circumvent the above limitations, but they have not been exhaustively explored yet. This paper investigates diverse LLMs and learning-based transpilers for automated code translation tasks, finding that: although certain LLMs have outperformed current transpilers, they still have some accuracy issues, where most of the failures are induced by a lack of comprehension of source programs (38.51%), missing clear instructions on I/O types in translation (14.94%), and ignoring discrepancies between source and target programs (41.38%). Enlightened by the above findings, we further propose UniTrans , a Unified code Translation framework, applicable to various LLMs, for unleashing their power in this field. Specifically, UniTrans first crafts a series of test cases for target programs with the assistance of source programs. Next, it harnesses the above auto-generated test cases to augment the code translation and then evaluate their correctness via execution. Afterward, UniTrans further (iteratively) repairs incorrectly translated programs prompted by test case execution results. Extensive experiments are conducted on six settings of translation datasets between Python, Java, and C++. Three recent LLMs of diverse sizes, including GPT-3.5 and LLaMA-13B/7B, are tested with UniTrans , and all achieve substantial improvements in terms of computational accuracy and exact match accuracy among almost all translation settings, showing the universal effectiveness of UniTrans in practice.
Förderung
- National Key R&D Program
- National Natural Science Foundation of China
- Shandong Province Overseas Outstanding Youth Fund
- City University of Hong Kong
- Key Program of Hubei
Zitieren
APA
Yang, Z., Liu, F., Yu, Z., Keung, J. W., Li, J., Liu, S., Hong, Y., Ma, X., Jin, Z., & Li, G. (2024). Exploring and Unleashing the Power of Large Language Models in Automated Code Translation. Proceedings of the ACM on Software Engineering, 1(FSE), 1585-1608. https://doi.org/10.1145/3660778
Vancouver
Yang Z, Liu F, Yu Z, Keung JW, Li J, Liu S, et al. Exploring and Unleashing the Power of Large Language Models in Automated Code Translation. Proceedings of the ACM on Software Engineering. 2024;1(FSE):1585-1608. doi:10.1145/3660778
BibTeX
@article{Yang2024,
author = {Yang, Zhen and Liu, Fang and Yu, Zhongxing and Keung, Jacky Wai and Li, Jia and Liu, Shuo and Hong, Yifan and Ma, Xiaoxue and Jin, Zhi and Li, Ge},
title = {Exploring and Unleashing the Power of Large Language Models in Automated Code Translation},
journal = {Proceedings of the ACM on Software Engineering},
year = {2024},
volume = {1},
number = {FSE},
pages = {1585--1608},
doi = {10.1145/3660778}
}
EUVIMED