Xiangyang Jia, male, graduated from Wuhan University in June 2008 with a PhD degree in Software Engineering, and stayed in Wuhan University to engage in software engineering research, teaching and software development, and went to Politecnico di Milano in 2014 for one year to study in the DEEPSE group of Prof. Carlo Ghezzi (winner of ACM SIGSOFT Outstanding Research Award).

He has received funding from Natural Science Foundation of China (NSFC), and participated in many national research projects, such as the National Key Research and Development Program, the 973 project, the 863 project, etc. He has research cooperation projects with state-owned enterprises such as CECC, Aerospace Science and Technology Group, and Hubei Grid, as well as with Internet giants such as Huawei, Oppo, and so on. He has published many papers in high level international conferences and journals recognized by CCF, and has had many collaborations and exchanges with scholars and teams in the field of soft engineering at home and abroad.

He has rich experience in architecture design and software development. As the R&D leader and architect, he has organized the R&D of more than 10 large-scale software platforms. The “Service-oriented Information Integration Platform SOAWARE” has been deployed in more than 40 electric power enterprises nationwide; the “Comprehensive Disaster Reduction Information Service Platform” has been deployed in the Disaster Reduction Center of the Ministry of Civil Affairs of the People's Republic of China. The “Spatial Data Intelligent Governance Platform” has been deployed and used in Beijing, Hefei, Hainan, Chengdu and other cities; the “AI Open Platform” has been deployed in Beijing and Chongqing.

Research Team

I have a research team and a full-time development team. The two teams cowork to finding academic problems in engineering and solve engineering challenges in lab. 

The research team consists of graduate and undergraduate students, and the main research interests include: 

- AI for Software Engineering: With the rapid development of LLMs, the software as well as its development are experiencing revolutionary changes. This direction focuses on the application of LLMs in software development, including code generation, test automation, defect detection, code comprehension, and document generation. The current work in progress is a QA system for code repositorys, to help developer understanding project features and architectures, as well as generating code that meets project requirements.

- AI security: the popularity of AI and LLMs highlights their security issues. The current works include adversarial attack and defense of visual models, backdoor attack and defense of large model intelligences, and image forgery detection and forensics. 

The development team mainly composed of full-time developers, and students can participate to some extent. The main works currently includes: 

- AI open platform:  a platform for AI model training, inference and application. It includs five sub-platforms:  intelligent data workshop, self-service training workshop, AI model marketplace, AI model evaluation center, and Lab of AI agents.

- Spatio-Temporal Intelligence Platform: Supports aggregation, search, interpretation and distribution of various types of spatio-temporal data, such as remote sensing images, drone images, vectors, trajectories, and so on. It supports various capabilities such as QA-style data search, object detection, segmentation, and data recommendation. 

- AI applications: including all kinds of Agents (Q&A on knowledge base, work assistant, copywriting assistant, etc.), in-vehicle smart boxes, drone inspection and inventory, industrial defect detection, etc.

Publications

https://dblp.org/pid/82/2706.html