Recent Updates
- Mar, 2024 Our papers "API Misuse Detection via Probabilistic Graphical Model" and "Model-less Is The Best Model: Generating Pure Code Implementations to Replace On-device DL Models" have been accepted by ISSTA 2024 research paper track.
- Mar, 2024 Our SANER'24 paper "Investigating and Detecting Silent Bugs in PyTorch Programs" won IEEE TCSE Distinguished Paper Award.
- Feb, 2024 Our ICSE'24 paper "Modularizing while Training: A New Paradigm for Modularizing DNN Models" won ACM SIGSOFT Distinguished Paper Award.
- Jan, 2024 our paper "ProveNFix: Temporal Property guided Program Repair" has been accepted by the FSE'24 research papers track.
- Dec, 2023 our paper "Investigating White-Box Attacks for On-Device Models" has been accepted by the ICSE'24 research papers track.
- Nov, 2023 our paper "Reusing Convolutional Neural Network Models through Modularization and Composition" has been accepted by TOSEM as research paper.
- Aug, 2023 our paper "Automated Fixing of Web UI Tests via Iterative Element Matching" has been accepted by the ASE'23 research papers track.
- Jun, 2023 our paper "Modularizing while Training: A New Paradigm for Modularizing DNN Models" has been accepted by the ICSE'24 research papers track.
- May, 2023 our paper "ModelObfuscator: Obfuscating Model Information to Protect Deployed ML-based Systems" has been accepted by the ISSTA'23 technical papers track.
- Dec, 2022 Binhang's paper Reusing Deep Neural Network Models through Model Re-engineering" has been accepted by the ICSE'23 technical track.
Research
Research Directions
To improve software development efficiency and save cost, when developing software, many enterprises rely on third-party software, which constitutes the software supply chain (SSC). Due to the dependency relations among software, vulnerabilities in SSC may cause more serious security threats than independent software systems. This poses new challenges for ensuring software security. Our research mainly focuses on designing novel intelligent software engineering and program analysis techniques, including but not limited to software dependency analysis, software vulnerability detection, vulnerability repair and etc. The ultimate goal is to increase the security of software supply chain and software ecosystem.
Usability and Robustness of DNN via SE4AI
Due to poor interpretability, a large number of parameters and data requirements, and poor reliability, the AI model, as the core of intelligent software systems, suffers from poor reusability, high testing overhead, and high security risks in development, testing, and deployment. AI models are considered as "Software 2.0". In this direction, we target the above problems from the perspective of software engineering. we aim to apply software engineering techniques/notions to AI model engineering to improve the models' usability and robustness.
AI/Crowd Intelligence
Crowd intelligence aims at achieving powerful intelligence through aggregating diverse contributions from many heterogeneous individuals, which is a typical form of AI. As a matter of fact, the sucess of many AI tehcnologies highly relies on large-scale datasets that are often built through crowd intelligence. We study various perspectives of crowd intelligence including theories, techniques and applications. Research topics include, but are not limited to, crowdsourcing, federated learning, human-in-the-loop AI and open source.
Selected Research Papers
-
API Misuse Detection via Probabilistic Graphical Model
Yunlong Ma, Wentong Tian, Xiang Gao, Hailong Sun, Li Li.International Symposium on Software Testing and Analysis (ISSTA), 2024.
-
Modularizing while Training: A New Paradigm for Modularizing DNN Models
Binhang Qi, Hailong Sun, Hongyu Zhang, Ruobing Zhao, Xiang Gao.
International Conference on Software Engineering (ICSE), 2024.
-
Investigating and Detecting Silent Bugs in PyTorch Programs
Shuo Hong, Hailong Sun, Xiang Gao#, Shin Hwei Tan.
International Conference on Software Analysis, Evolution and Reengineering (SANER), 2024.
-
Automated Repair of Programs from Large Language Models
Zhiyu Fan, Xiang Gao#, Martin Mirchev, Abhik Roychoudhury, Shin Hwei Tan.
International Conference on Software Engineering (ICSE) 2023, 2023.
-
Reusing Deep Neural Network Models through Model Re-engineering
Binhang Qi, Hailong Sun, Xiang Gao, Hongyu Zhang, Zhaotian Li, Xudong Liu.
The 45th International Conference on Software Engineering (ICSE), 2023.
-
Template-based Neural Program Repair
Xiangxin Meng, Xu Wang, Hongyu Zhang, Hailong Sun, Xudong Liu, Chunming Hu.The 45th International Conference on Software Engineering (ICSE), 2023.
-
A Collaboration-Aware Approach to Profiling Developer Expertise with Cross-Community Data
Xiaotao Song, Jiafei Yan, Yuexin Huang, Hailong Sun, Hongyu Zhang.The 22nd IEEE International Conference on Software Quality, Reliability, and Security (QRS), 2022.
-
Improving Fault Localization and Program Repair with Deep Semantic Features and Transferred Knowledge
Xiangxin Meng, Xu Wang, Hongyu Zhang, Hailong Sun, Xudong Liu.The 44th International Conference on Software Engineering (ICSE), 2022.
-
Retrieval-based Neural Source Code Summarization
Jian Zhang, Xu Wang, Hongyu Zhang, Hailong Sun, Xudong Liu.The 42nd International Conference on Software Engineering (ICSE), 2020.
ACM SIGSOFT Distinguished Paper Award
IEEE TCSE Distinguished Paper Award