arrow
Return

ICAM-Trans: Implicit-Context-Aware Multi-Agent Framework for Function-Level Code Translation

delete2026-01-01
delete0
PRE
AI
R
Ruolin Chen
王
王杰新 (Jiexin Wang)
L
Liuwen Cao
蔡毅 cover
蔡毅 (Yi Cai)
M
Mengchen Zhao
DOI:10.1145/3810952delete
deleteOriginal
deleteShare
deleteSave
Abstract

Abstract

En 中文
With the rapid advancement of large language models (LLMs), code translation has become a critical yet challenging task in software engineering. Existing evaluations of this task remain unreliable due to two key limitations: the absence of specialized function-level code translation benchmarks and flawed evaluation pipelines. Specifically, current pipelines lack post-processing mechanisms to extract target code from LLMs’ inherently unstable output formats, directly resulting in non-executable translations and false negatives in evaluation. To address these issues, we introduce FuncTransEval, an enhanced multilingual benchmark. It comprises 400×4 functions across four programming languages (Python, C++, Java, and JavaScript) and 12 bidirectional translation pairs, each paired with executable unit tests for fine-grained functional validation. Using FuncTransEval, we systematically assess LLMs’ true capabilities in code translation and identify their core limitations: inadequate understanding of the implicit information such as function logic, type specifications, and cross-language differences. To mitigate these shortcomings, we propose ICAM-Trans (Implicit-Context-Aware Multi-Agent Framework for function-level code translation). ICAM-Trans is a hierarchical, autonomous multi-agents architecture that explicitly uncovers and leverages these implicit contextual information during translation. It employs a Translation Orchestrate Agent (TOA) to autonomously coordinate four specialized Context Analysis Agents (CAAs), which conduct pre-translation analyses on different aspects. Insights from these analyses guide a Context-Aware Translation Agent (CTA) to generate semantically faithful target code. Experiments on FuncTransEval demonstrate that ICAM-Trans consistently outperforms strong baselines, validating its effectiveness in achieving high-fidelity and interpretable function-level code translation.

Journal

A
ACM Transactions on Software Engineering and Methodology
IF:
6.2
Papers:
1.4K
Citations:
4.5K

Organization

South China University of Technology cover
South China University of Technology
Scholars:
6.0K
Papers: 2.0K
Citations: 4.4W
Cited Papers

Cited Papers

New Frontiers of Reverse Engineering
err2007-05-01
err0
PREAI
errGerardo Canfora; Massimiliano Di Penta
errShare
errSave
Source-to-source translation
err1980-11-01
err0
errOAAI
errPaul F. Albrecht; Philip E. Garrison; Susan L. Graham; Robert H. Hyerle; Patricia Ip; Bernd Krieg-Brückner
errShare
errSave
Using machine translation for converting Python 2 to Python 3 code
err
IF0
err2015-10-29
err0
errOAAI
errKaran Aggarwal; Mohammad Salameh; Abram Hindle
errShare
errSave
ClassEval-T: Evaluating Large Language Models in Class-Level Code Translation
err2025-06-22
err0
PREAI
errXue,Pengyu; Wu,Linhao; Yang,Zhen; Wang,Chengyi; Li,Xiang; Zhang,Yuxiang; Li,Jia; Jin,Ruikai; Pei,Yifei; Shen,Zhaoyan; Lyu,Xiran; Keung,Jacky Wai
errShare
errSave
CodeGeeX: A Pre-Trained Model for Code Generation with Multilingual Benchmarking on HumanEval-X
err2023-08-04
err0
errOAAI
errQinkai Zheng; Xiao Xia; Xu Zou; Yuxiao Dong; Shan Wang; Yufei Xue; Lei Shen; Zihan Wang; Andi Wang; Yang Li; Teng Su; Zhilin Yang; Jie Tang
errShare
errSave
RepoTransBench: A Real-World Multilingual Benchmark for Repository-Level Code Translation
err2025-12-17
err0
PREAI
errYanli Wang; Yanlin Wang; Suiquan Wang; Daya Guo; Jiachi Chen; John Grundy; Xilin Liu; Yuchi Ma; Mingzhi Mao; Hongyu Zhang; Zibin Zheng
errShare
errSave
researcher View more