Wenkang Zhong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Wenkang Zhong has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Chuanyi Li
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China
6 shared publications - Bin Luo
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China
5 shared publications - Jidong Ge
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China
5 shared publications - Vincent Ng
Human Language Technology Research Institute, University of Texas at Dallas, Richardson, TX, USA
4 shared publications - Zongwen Shen
State Key Laboratory for Novel Software Technology, Nanjing University, China
3 shared publications - Qingyuan Li
National Key Laboratory for Novel Software Technology at Nanjing University, China
2 shared publications
Latest Publications
- Open Access
- 3 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Towards High-Quality Neural Program Repair: Learning from Positive Code Changes
Automated Program Repair (APR) seeks to fix bugs without human intervention, aiming to free developers from the tedious and time-consuming debugging process. In recent years, with the advancement of deep learning techniques, Neural Program Repair (NPR) has emerged as a promising research direction for APR. NPR tools have shown significant progress, especially in terms of the number of bugs correctly fixed across well-established benchmarks. However, recent studies have raised concerns regarding the reliability of NPR tools, revealing that tools trained on bug fix data also perform well in fault injection tasks. This suggests that existing NPR tools may not be genuinely fixing bugs, but rather simply making code modifications. As a result, the patches generated by NPR tools have a high likelihood of introducing new bugs, posing a potential risk to software reliability. In this study, we investigate the underlying reasons behind this phenomenon. We begin by conducting an empirical study to investigate different types of code change operations. By performing clustering analysis on a large number of code change samples, we categorize code changes into three distinct groups based on the nature of fixing code change patterns: positive, mixed, and negative. Next, we conduct experiments to investigate how different types of code changes influence NPR tools’ performance. We utilize the code change templates derived from the manual analysis of these clusters to filter the existing dataset and retrain NPR models using this refined dataset. Experimental results show that the retrained NPR models outperform their non-retrained counterparts in terms of precision, and the generated patches contain fewer negative code changes. Based on the results, we discuss insights into the limitations of previous NPR studies and hope to provide researchers with a new perspective for enhancing their reliability by improving dataset quality.
- 3 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Open AccessTowards High-Quality Neural Program Repair: Learning from Positive Code Changes
Automated Program Repair (APR) seeks to fix bugs without human intervention, aiming to free developers from the tedious and time-consuming debugging process. In recent years, with the advancement of deep learning techniques, Neural Program Repair (NPR) has emerged as a promising research direction for APR. NPR tools have shown significant progress, especially in terms of the number of bugs correctly fixed across well-established benchmarks. However, recent studies have raised concerns regarding the reliability of NPR tools, revealing that tools trained on bug fix data also perform well in fault injection tasks. This suggests that existing NPR tools may not be genuinely fixing bugs, but rather simply making code modifications. As a result, the patches generated by NPR tools have a high likelihood of introducing new bugs, posing a potential risk to software reliability. In this study, we investigate the underlying reasons behind this phenomenon. We begin by conducting an empirical study to investigate different types of code change operations. By performing clustering analysis on a large number of code change samples, we categorize code changes into three distinct groups based on the nature of fixing code change patterns: positive, mixed, and negative. Next, we conduct experiments to investigate how different types of code changes influence NPR tools’ performance. We utilize the code change templates derived from the manual analysis of these clusters to filter the existing dataset and retrain NPR models using this refined dataset. Experimental results show that the retrained NPR models outperform their non-retrained counterparts in terms of precision, and the generated patches contain fewer negative code changes. Based on the results, we discuss insights into the limitations of previous NPR studies and hope to provide researchers with a new perspective for enhancing their reliability by improving dataset quality.
- 28 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Software Engineering
- Vol. PP (00985589) , 1-16
P-NPR: Practical Neural Program Repair via Learning to Ensemble
The rapid emergence of Large Language Model (LLM)-based program repair tools has significantly advanced the potential of automated program repair. Nevertheless, traditional Neural Program Repair (NPR) tools continue to offer unique value in resource-constrained or latency-sensitive scenarios due to their lightweight nature. Over 50 such NPR tools have been designed and proven effective at fixing various bugs using only the buggy code as input. Recent studies further demonstrate that combining these tools through ensemble strategies can enhance overall repair performance by leveraging their complementary strengths. However, existing ensemble approaches still suffer from two critical limitations: (1) they rely on manual feature summarization, requiring labor-intensive labeling of bug characteristics, and (2) they ignore the common “all-tools-fail” scenario (e.g., even a set of 15 state-of-the-art tools leaves over 66% of Defects4J bugs unfixed), which incurs unnecessary computational overhead. To address these issues, we propose Practical Neural Program Repair (P-NPR), a novel ensemble strategy that employs multilabel classification to automatically learn to characterize bugs and introduces a dedicated “AllFailure” label to explicitly sidestep futile repair attempts. Experimental results show that P-NPR outperforms the state-of-the-art ensemble strategy (i.e., P-EPR) by over 65% in Top1-accuracy, while reducing runtime by up to 45% and GPU memory usage by up to 23% across evaluated datasets.
- 28 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Software Engineering
- Vol. PP (00985589) , 1-16
P-NPR: Practical Neural Program Repair via Learning to Ensemble
The rapid emergence of Large Language Model (LLM)-based program repair tools has significantly advanced the potential of automated program repair. Nevertheless, traditional Neural Program Repair (NPR) tools continue to offer unique value in resource-constrained or latency-sensitive scenarios due to their lightweight nature. Over 50 such NPR tools have been designed and proven effective at fixing various bugs using only the buggy code as input. Recent studies further demonstrate that combining these tools through ensemble strategies can enhance overall repair performance by leveraging their complementary strengths. However, existing ensemble approaches still suffer from two critical limitations: (1) they rely on manual feature summarization, requiring labor-intensive labeling of bug characteristics, and (2) they ignore the common “all-tools-fail” scenario (e.g., even a set of 15 state-of-the-art tools leaves over 66% of Defects4J bugs unfixed), which incurs unnecessary computational overhead. To address these issues, we propose Practical Neural Program Repair (P-NPR), a novel ensemble strategy that employs multilabel classification to automatically learn to characterize bugs and introduces a dedicated “AllFailure” label to explicitly sidestep futile repair attempts. Experimental results show that P-NPR outperforms the state-of-the-art ensemble strategy (i.e., P-EPR) by over 65% in Top1-accuracy, while reducing runtime by up to 45% and GPU memory usage by up to 23% across evaluated datasets.
- Open Access
- 12 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (4) , 1-36
IMPACT: Identifying and Classifying Multiple Sourced and Categorized Self-Admitted Technical Debts
Self-Admitted Technical Debt (SATD) refers to sub-optimal solutions deliberately introduced to accelerate the software development process, often at the expense of software maintainability and sustainability. Therefore, timely identification and repayment of the SATD is critical for the software system. As exploration deepens, it is found that effectively prioritizing the repayment of SATD with more significant impacts on software quality requires not only identifying SATD but also further classifying it. However, existing SATD identification and classification approaches face the following challenges: (1) SATDs originate from diverse sources. Code comments are a widespread source, but recent research has revealed that SATDs can originate from other sources, such as pull requests, issues, and commit messages. Nonetheless, existing approaches primarily target code comments, lacking the capability to analyze SATDs from other sources effectively. (2) SATDs fall into diverse categories. Nonetheless, existing SATD classification approaches fail to address all SATD categories comprehensively and show inadequate performance. (3) Imbalance of existing SATD datasets. Real-world SATD data are scarce, making dataset collection challenging. Moreover, SATD distribution across different sources is uneven, further complicating the construction of high-quality datasets. To alleviate these challenges, this article presents an SATD identification and classification framework named IMPACT . First, IMPACT employs ChatGPT to construct an augmented dataset. Subsequently, it utilizes a pipeline with two fine-tuned language models of different parameter sizes to identify and classify SATD separately. To evaluate the effectiveness of IMPACT, we compare it with three state-of-the-art SATD classification methods and its two foundation models. Experimental results demonstrate that IMPACT outperforms state-of-the-art methods by a large margin, and even surpasses its foundation model GLM-4-9B-Chat. It achieves the optimal average F1 score of 0.697 on the source of pull requests, the most challenging data source. Moreover, experiments on the cross-project test set show that IMPACT demonstrates strong generalizability on unseen project data.
- 12 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (4) , 1-36
Open AccessIMPACT: Identifying and Classifying Multiple Sourced and Categorized Self-Admitted Technical Debts
Self-Admitted Technical Debt (SATD) refers to sub-optimal solutions deliberately introduced to accelerate the software development process, often at the expense of software maintainability and sustainability. Therefore, timely identification and repayment of the SATD is critical for the software system. As exploration deepens, it is found that effectively prioritizing the repayment of SATD with more significant impacts on software quality requires not only identifying SATD but also further classifying it. However, existing SATD identification and classification approaches face the following challenges: (1) SATDs originate from diverse sources. Code comments are a widespread source, but recent research has revealed that SATDs can originate from other sources, such as pull requests, issues, and commit messages. Nonetheless, existing approaches primarily target code comments, lacking the capability to analyze SATDs from other sources effectively. (2) SATDs fall into diverse categories. Nonetheless, existing SATD classification approaches fail to address all SATD categories comprehensively and show inadequate performance. (3) Imbalance of existing SATD datasets. Real-world SATD data are scarce, making dataset collection challenging. Moreover, SATD distribution across different sources is uneven, further complicating the construction of high-quality datasets. To alleviate these challenges, this article presents an SATD identification and classification framework named IMPACT . First, IMPACT employs ChatGPT to construct an augmented dataset. Subsequently, it utilizes a pipeline with two fine-tuned language models of different parameter sizes to identify and classify SATD separately. To evaluate the effectiveness of IMPACT, we compare it with three state-of-the-art SATD classification methods and its two foundation models. Experimental results demonstrate that IMPACT outperforms state-of-the-art methods by a large margin, and even surpasses its foundation model GLM-4-9B-Chat. It achieves the optimal average F1 score of 0.697 on the source of pull requests, the most challenging data source. Moreover, experiments on the cross-project test set show that IMPACT demonstrates strong generalizability on unseen project data.
- Open Access
- 28 August 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (8) , 1-62
Patch Generation in APR: A Survey from the Perspectives of Utilizing LLMs and Using APR-Specific Information
Automated Program Repair (APR) is a crucial task in software development and maintenance, aiming to patch software bugs automatically without human intervention. The rise of Large Language Models (LLMs) has significantly advanced APR. However, as researchers delve deeper into APR as a downstream task for LLMs, a key challenge remains: how to effectively integrate APR-specific information to complement LLMs capabilities. This survey revisits 124 existing APR studies, most from 2021 to 2024, from two perspectives: Utilizing LLMs and APR-specific Information. First, we define the concept of APR-specific information. Then, we summarize techniques for utilizing LLMs in patch generation in four dimensions. Next, we explore critical factors influencing the effectiveness of generating patches by LLMs, such as prompting context and fine-tuning configurations. After that, we focus on the evaluation of generated patches, including benchmarks, reasoning cost scaling, etc. Furthermore, we distill all APR-specific information (e.g., bug-fix pairs and error messages), highlighting their unique importance and features. Finally, we comprehensively outline challenges, limitations, and potential future research directions in APR in the LLM era. To conclude, by adopting the two perspectives, we aim to provide valuable insights into mining and leveraging APR-specific information in utilizing LLMs process for the APR community.
- 28 August 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (8) , 1-62
Open AccessPatch Generation in APR: A Survey from the Perspectives of Utilizing LLMs and Using APR-Specific Information
Automated Program Repair (APR) is a crucial task in software development and maintenance, aiming to patch software bugs automatically without human intervention. The rise of Large Language Models (LLMs) has significantly advanced APR. However, as researchers delve deeper into APR as a downstream task for LLMs, a key challenge remains: how to effectively integrate APR-specific information to complement LLMs capabilities. This survey revisits 124 existing APR studies, most from 2021 to 2024, from two perspectives: Utilizing LLMs and APR-specific Information. First, we define the concept of APR-specific information. Then, we summarize techniques for utilizing LLMs in patch generation in four dimensions. Next, we explore critical factors influencing the effectiveness of generating patches by LLMs, such as prompting context and fine-tuning configurations. After that, we focus on the evaluation of generated patches, including benchmarks, reasoning cost scaling, etc. Furthermore, we distill all APR-specific information (e.g., bug-fix pairs and error messages), highlighting their unique importance and features. Finally, we comprehensively outline challenges, limitations, and potential future research directions in APR in the LLM era. To conclude, by adopting the two perspectives, we aim to provide valuable insights into mining and leveraging APR-specific information in utilizing LLMs process for the APR community.
- Open Access
- 30 December 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 34 (1) , 1-35
Benchmarking and Categorizing the Performance of Neural Program Repair Systems for Java
Recent years have seen a rise in Neural Program Repair (NPR) systems in the software engineering community, which adopt advanced deep learning techniques to automatically fix bugs. Having a comprehensive understanding of existing systems can facilitate new improvements in this area and provide practical instructions for users. However, we observe two potential weaknesses in the current evaluation of NPR systems: ① published systems are trained with varying data, and ② NPR systems are roughly evaluated through the number of totally fixed bugs. Questions such as what types of bugs are repairable for current systems cannot be answered yet. Consequently, researchers cannot make target improvements in this area and users have no idea of the real affair of existing systems. In this article, we perform a systematic evaluation of the existing nine state-of-the-art NPR systems. To perform a fair and detailed comparison, we (1) build a new benchmark and framework that supports training and validating the nine systems with unified data and (2) evaluate re-trained systems with detailed performance analysis, especially on the effectiveness and the efficiency. We believe our benchmark tool and evaluation results could offer practitioners the real affairs of current NPR systems and the implications of further facilitating the improvements of NPR.
- 30 December 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 34 (1) , 1-35
Open AccessBenchmarking and Categorizing the Performance of Neural Program Repair Systems for Java
Recent years have seen a rise in Neural Program Repair (NPR) systems in the software engineering community, which adopt advanced deep learning techniques to automatically fix bugs. Having a comprehensive understanding of existing systems can facilitate new improvements in this area and provide practical instructions for users. However, we observe two potential weaknesses in the current evaluation of NPR systems: ① published systems are trained with varying data, and ② NPR systems are roughly evaluated through the number of totally fixed bugs. Questions such as what types of bugs are repairable for current systems cannot be answered yet. Consequently, researchers cannot make target improvements in this area and users have no idea of the real affair of existing systems. In this article, we perform a systematic evaluation of the existing nine state-of-the-art NPR systems. To perform a fair and detailed comparison, we (1) build a new benchmark and framework that supports training and validating the nine systems with unified data and (2) evaluate re-trained systems with detailed performance analysis, especially on the effectiveness and the efficiency. We believe our benchmark tool and evaluation results could offer practitioners the real affairs of current NPR systems and the implications of further facilitating the improvements of NPR.
- Open Access
- 6 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Practical Program Repair via Preference-based Ensemble Strategy
To date, over 40 Automated Program Repair (APR) tools have been designed with varying bug-fixing strategies, which have been demonstrated to have complementary performance in terms of being effective for different bug classes. Intuitively, it should be feasible to improve the overall bug-fixing performance of APR via assembling existing tools. Unfortunately, simply invoking all available APR tools for a given bug can result in unacceptable costs on APR execution as well as on patch validation (via expensive testing). Therefore, while assembling existing tools is appealing, it requires an efficient strategy to reconcile the need to fix more bugs and the requirements for practicality. In light of this problem, we propose a Preference-based Ensemble Program Repair framework (P-EPR), which seeks to effectively rank APR tools for repairing different bugs. P-EPR is the first non-learning-based APR ensemble method that is novel in its exploitation of repair patterns as a major source of knowledge for ranking APR tools and its reliance on a dynamic update strategy that enables it to immediately exploit and benefit from newly derived repair results. Experimental results show that P-EPR outperforms existing strategies significantly both in flexibility and effectiveness.
- 6 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessPractical Program Repair via Preference-based Ensemble Strategy
To date, over 40 Automated Program Repair (APR) tools have been designed with varying bug-fixing strategies, which have been demonstrated to have complementary performance in terms of being effective for different bug classes. Intuitively, it should be feasible to improve the overall bug-fixing performance of APR via assembling existing tools. Unfortunately, simply invoking all available APR tools for a given bug can result in unacceptable costs on APR execution as well as on patch validation (via expensive testing). Therefore, while assembling existing tools is appealing, it requires an efficient strategy to reconcile the need to fix more bugs and the requirements for practicality. In light of this problem, we propose a Preference-based Ensemble Program Repair framework (P-EPR), which seeks to effectively rank APR tools for repairing different bugs. P-EPR is the first non-learning-based APR ensemble method that is novel in its exploitation of repair patterns as a major source of knowledge for ranking APR tools and its reliance on a dynamic update strategy that enables it to immediately exploit and benefit from newly derived repair results. Experimental results show that P-EPR outperforms existing strategies significantly both in flexibility and effectiveness.