教師簡介：

黃袁，博士/博士后，副教授，博士生導(dǎo)師。主持國家重點研發(fā)計劃子任務(wù)，國家自然科學(xué)基金青年項目，博士后面上項目（一等資助），廣東省基金面上項目等。以項目骨干身份參與國家重點研發(fā)計劃，國家基金重點項目，廣東省重點研發(fā)計劃等多個項目。2017年于必贏3003no1線路檢測中心獲得博士學(xué)位，從事軟件工程相關(guān)研究，重點關(guān)注軟件缺陷與代碼智能等研究方向。近年來在IEEE Transactions on Software Engineering, ACM Transactions on Software Engineering and Methodology, IEEE Transactions on Services Computing, FSE/ESEC, ASE, ICSE等軟件工程領(lǐng)域CCF A/B期刊及會議上發(fā)表論文40余篇。同時擔(dān)任多個國際期刊和會議的審稿人。獲“2024優(yōu)秀碩士學(xué)位論文指導(dǎo)教師”獎項。

 

研究領(lǐng)域:

主要研究方向包括：

-代碼缺陷定位與修復(fù)：關(guān)注多種類型的缺陷檢測與修復(fù)，包括緩沖區(qū)溢出漏洞、即時缺陷定位、智能合約漏洞、測試代碼失效缺陷、代碼文檔過時缺陷等。涉及技術(shù)：代碼靜態(tài)分析、思維鏈CoT、深度學(xué)習(xí)、區(qū)塊鏈、智能合約等。

-代碼大模型幻覺檢測與消除：針對AI-Generated Code，利用程序分析、統(tǒng)計分析等手段，檢測其中是否存在幻覺問題；同時從代碼大模型的輸入及構(gòu)造機(jī)理入手，限制大模型生成幻覺代碼。

-代碼及其文檔自動生成：關(guān)注從需求描述（自然語言）自動生成源代碼，以及從代碼（源代碼或字節(jié)碼）自動生成自然語言描述的軟件文檔。涉及技術(shù)：大模型、微調(diào)Fine-tuning、自然語言處理技術(shù)、代碼靜態(tài)分析、反匯編等。

 

招生：歡迎對智能化軟件工程、大模型、區(qū)塊鏈有興趣的研究生及本科生同學(xué)們加入課題組。組內(nèi)氣氛融洽，每年組織多次團(tuán)建活動，大部分研究生及本科生同學(xué)進(jìn)組后可快速獲得成長并發(fā)表CCF A/B期刊或會議（見下文“學(xué)生培養(yǎng)情況”）。每年招收2-3名碩士研究生、與 1-2 名博士研究生（包括直博、碩博連讀）；歡迎對科研有濃厚興趣的本科生加入課題組。

郵箱：huangyuan5@mail.sysu.edu.cn

 

更多最新信息請參見個人主頁：https://huangyuan-hy.github.io/

 

最新動態(tài)：

  • December 17, 2024, "Towards improving the performance of comment generation models by using bytecode information" was accepted at IEEE Transactions on Software Engineering (CCF A). The bytecode information is introduced to improve the accruccy of the comment generation models in this study.
  • November 27，2024，"TG-CUP: A Transformer and GNN Based Multi-Modal Comment Updating Method" was accepted at ACM Transactions on Software Engineering and Methodology (CCF A). We propose a new outdated comment update model TG-CUP in this study.
  • August 7, 2024, "An Empirical Study on Learning-based Techniques for Explicit and Implicit Commit Messages Generation" was accepted at ASE 2024（CCF A）. This work studies the performance of the learning-based models on the task of explicit and implicit commit message generation.
  • June 21, 2024, 祝賀小組3位碩士生（徐福仁、黃進(jìn)波、王榕）+1位本科生（李陽姿）順利畢業(yè)。其中，徐福仁和李陽姿分獲“優(yōu)秀畢業(yè)生”稱號，徐福仁獲“優(yōu)秀碩士學(xué)位論文”（全院唯一）。
  • April 15, 2024, 發(fā)布以太坊并行重放平臺EthReplayer，該平臺允許使用真實的Transactions測試智能合約缺陷修復(fù)、智能合約Gas優(yōu)化、智能合約Gas預(yù)估等任務(wù)。URL：https://github.com/gastool/EthReplayer/tree/replay2
  • April 3, 2024, "Ethereum Transaction Replay Platform Based on State-wise Account Input Data" was accepted at IEEE Transactions on Services Computing (CCF A). In this study we propose an efficient transaction replay platform, EthReplayer, which can not only replay the historical transactions of Ethereum quickly and faithfully but also realize the modification of transactions to achieve the purpose of testing with actual transactions.
  • March 5, 2024, "Generative Software Engineering（生成式軟件工程）" is posted on the arXiv. This paper presents a comprehensive literature review of generative tasks in software engineering by applying large language models (LLMs). URL: https://arxiv.org/abs/2403.02583
  • February 17, 2024,  "Do Code Summarization Models Process Too Much Information? Function Signature May Be All What Is Needed" was accepted at ACM Transactions on Software Engineering and Methodology (CCF A). This study found that there is no significant difference in BLEU value between comments generated using the function signature and those generated using the entire function as input.
  • January 23, 2024,  "JIT-Smart: A Multi-Task Learning Framework for Just-In-Time Defect Prediction and Localization" was accepted at FSE/ESEC 2024（CCF A）. This work was completed by Furen Xu (碩士生徐福仁). Unlike traditional methods, which usually model just-in-time defect prediction and localization as a two-stage task, JIT-Smart models the commit-level prediction task and the line-level localization task in a unified way. In the prediction phase, JIT-Smart outperforms 7 SOTA methods, and in the localization phase, JIT-Smart outperforms 3 SOTA methods. The model is open source: https://github.com/JIT-A/JIT-Smart. （在CCF軟件工程專委會頂級國際會議之早鳥交流會上匯報相關(guān)研究）
  • January 11, 2024,  "ESGen: Commit Message Generation Based on Edit Sequence of Code Change" was accepted at ICPC 2024 (CCF B). This work was completed by Yangzi Li (本科生李陽姿), an undergraduate student in our research group.
  • See More：https://huangyuan-hy.github.io/

 

科研項目:

  1. 2024-2026：（主持）國家重點研發(fā)計劃 參與單位 子任務(wù)
  2. 2024-2025：（主持）工信部電子五所重點實驗室課題
  3. 2023-2025:  （主持） 廣東省自然科學(xué)基金面上項目
  4. 2020-2022：（主持）國家青年科學(xué)基金項目
  5. 2018-2021：（主持）博士后面上項目，一等資助
  6. 2020-2022：（主持）高校基本科研業(yè)務(wù)費(fèi)青年教師培育項目
  7. 2018-2021：（項目骨干）國家重點研發(fā)計劃
  8. 2020-2022：（項目骨干）廣東省重點領(lǐng)域研發(fā)計劃
  9. 2020-2022：（項目骨干）國家基金重點項目

 

近年代表作：

  • Yuan Huang, Jinbo Huang, Xiangping Chen, Zibin Zheng, Towards improving the performance of comment generation models by using bytecode information, IEEE Transactions on Software Engineering (TSE), 2024.(CCF-A)
  • Yuan Huang, Jinyu Jiang, Xiapu Luo, Xiangping Chen, Zibin Zheng, and Gang Huang, Change-Patterns Mapping: A Boosting Way for Change Impact Analysis, IEEE Transactions on Software Engineering (TSE), 2022.(CCF-A)
  • Yuan Huang, Xiangping Chen, Kai Hong, Zibin Zheng, Code Review Knowledge Perception: Fusing Multi-Features for Salient-Class Location, IEEE Transactions on Software Engineering (TSE), 2021.(CCF-A)
  • Yuan Huang, Junhuai Shu, Xiangping Chen, Xiapu Luo, Zibin Zheng, A comparative study on method comment and inline comment, ACM Transactions on Software Engineering and Methodology (TOSEM), 2023.(CCF-A)
  • Ding Xi, Peng Rui, Chen Xiangping, Yuan Huang, Bian Jing, Zheng Zibin. Do Code Summarization Models Process Too Much Information? Function Signature May Be All What Is Needed. ACM Transactions on Software Engineering and Methodology, (TOSEM), 2024.(CCF-A)
  • Yinan Chen, Yuan Huang, Xiangping Chen, Zibin Zheng. TG-CUP: A Transformer and GNN Based Multi-Modal Comment Updating Method. ACM Transactions on Software Engineering and Methodology, (TOSEM), 2024.(CCF-A)
  • Hanyang Guo, Xiangping Chen, Yuan Huang, Yanlin Wang, Zibin Zheng. Snippet Comment Generation Based on Code Context Expansion. ACM Transactions on Software Engineering and Methodology, (TOSEM), 2023.(CCF-A)
  • Hanyang Guo, Xiangping Chen,Yingye Chen, Yuan Huang, Zibin Zheng. Smart Contract Code Repair Recommendation based on Reinforcement Learning and Multi-metric Optimization. ACM Transactions on Software Engineering and Methodology, (TOSEM), 2023.(CCF-A)
  • Yuan Huang, Rong Wang, Xiangping Chen, and Zibin Zheng. Ethereum Transaction Replay Platform Based on State-wise Account Input Data. IEEE Transactions on Services Computing(TSC), 2024.(CCF-A)
  • Xiangping Chen,Furen Xu, Yuan Huang, Neng Zhang, Zibin Zheng, " JIT-Smart: A Multi-Task Learning Framework for Just-In-Time Defect Prediction and Localization,"  In Proceedings of the 2024 ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, FSE/ESEC 2024. (CCF-A)
  • Zhiquan Huang,Yuan Huang,Xiangping Chen,Xiaocong Zhou,Changlin Yang,Zibin Zheng. An Empirical Study on Learning-based Techniques for Explicit and Implicit Commit Messages Generation, In Proceedings of the 39th IEEE/ACM International Conference on Automated Software Engineering (ASE 2024).(CCF-A)
  • Yuan Huang, Xinyu Hu, Xiangping Chen, Yingfei Xiong, and Zibin Zheng. Learning code context information to predict comment locations. IEEE Transactions on Reliability (TR), 2020.
  • Yuan Huang, Xingjian Liang, Xiapu Luo, Xiangping Chen, Zibin Zheng, Xiaocong Zhou, Reviewing Rounds Prediction for Code Patches, Empirical Software Engineering (EMSE), 2022.
  • Xiangping Chen, Peiyong Liao, Queping Kong, Yuan Huang, and Xiaocong Zhou. Studying Differentiated Code to Support Smart Contract Update. Empirical Software Engineering (EMSE), 2023
  • Xiangping Chen, Xing Hu, Yuan Huang, et al. Deep Learning-based Software Engineering: Progress, Challenges, and Opportunities. SCIENCE CHINA Information Sciences, 2024，doi：10.1007/s11432-023-4127-5

 

學(xué)生培養(yǎng)情況：

畢業(yè)碩士生：

  • 徐福仁, 2024屆, 發(fā)表論文：[FSE'24] [ICPC'22] [JSS], 所獲獎勵：[優(yōu)秀碩士學(xué)位論文][2024屆優(yōu)秀碩士畢業(yè)生][研究生一等獎學(xué)金][招聯(lián)金融獎學(xué)金][小米獎學(xué)金], 畢業(yè)去向：騰訊（算法崗）  
  • 黃進(jìn)波, 2024屆, 發(fā)表論文：[TSE] [ASEJ], 所獲獎勵：[研究生一等獎學(xué)金][招聯(lián)金融獎學(xué)金][小米獎學(xué)金], 畢業(yè)去向：騰訊（算法崗） 
  • 王榕, 2024屆, 發(fā)表論文：[TSC] [QRS'22], 所獲獎勵：[研究生二等獎學(xué)金], 畢業(yè)去向：創(chuàng)業(yè)（跨國） 

畢業(yè)本科生：

  • 李陽姿, 2024屆, 發(fā)表論文: [ICPC'24], 所獲獎勵:[2024屆優(yōu)秀本科畢業(yè)生][優(yōu)秀本科畢業(yè)論文], 畢業(yè)去向：保研浙大
  • 江金昱, 2022屆, 發(fā)表論文: [TSE], 畢業(yè)去向：保研復(fù)旦

 

學(xué)術(shù)服務(wù)：

  • IEEE Transactions on Software Engineering (TSE)
  • ACM Transactions on Software Engineering and Methodology (TOSEM)
  •  Information and Software Technology（IST）
  • Journal of Systems and Software (JSS)
  • Journal of Software: Evolution and Process (JSME)
  • Software Testing, Verification and Reliability (STVR)
  • Software Quality Journal (SQJ)
  • Engineering Applications of Artificial Intelligence (EAAI)
  • International Conference on Blockchain and Trustworthy Systems 2020
  • ACM Symposium on Information, Computer and Communications Security 2020
  • IEEE Conference on Communications and Network Security 2020