Obálka

MSCoT: Structured Chain-of-Thought Generation for Multiple Programming Languages

Uloženo v:
Vydáno v:The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Conference Proceedings (2025), p. 1-8
Hlavní autor: Jin, Naizhu
Další autoři: Li, Zhong, Zhang, Tian, Zeng, Qingkai
Vydáno:
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
Datasets
Programming languages
Performance enhancement
Multilingualism
Multiagent systems
Neural networks
Task complexity
Economic
On-line přístup:Citation/Abstract
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstrakt:Conference Title: 2025 International Joint Conference on Neural Networks (IJCNN)Conference Start Date: 2025 June 30Conference End Date: 2025 July 5Conference Location: Rome, ItalyWith the rapid development of code intelligence, the application of multiple programming languages is becoming increasingly widespread. However, most existing code generation models mainly focus on a single or a few programming languages, resulting in unsatisfactory performance in a multilingual environment. Chain-of-Thought (CoT) reasoning can significantly improve the performance of the model without the need for retraining or fine-tuning the code generation model by reasonably decomposing complex code generation tasks into multiple subtasks and gradually deriving solutions for each subtask. Nevertheless, the existing CoT generation methods mainly concentrate on Python code, and the performance on other programming languages remains unclear.To fill this gap, we first constructed a CoT generation dataset for 12 programming languages through multi-agent technology. On this basis, we proposed a CoT generation method MSCoT applicable to multiple programming languages. By introducing CoT into the code generation large model, the performance of the code generation large model in a multilingual environment can be improved. Through large-scale empirical research, we compared the generalization abilities of MSCoT and the existing CoT generation methods on multiple programming languages and proved the effectiveness of MSCoT for multiple programming languages. In addition, we also designed a human study to prove the quality of the CoT generated by MSCoT. Finally, we open-sourced the model and dataset of MSCoT to promote the research on CoT generation for multiple programming languages.
DOI:10.1109/IJCNN64981.2025.11228673
Zdroj:Science Database