In a collaborative effort with the Hong Kong University of Science and Technology, the Department of Management Information at our center has proposed a novel fine-grained summary generation scenario. This innovative approach generates sentence-level summaries by leveraging method-level summaries. Additionally, a differentiation-aware and keyword-guided fine-grained summary generation model has been developed to assist software engineering scientists in quickly understanding code and enhancing code intelligence.

The research introduces a Keyword-Guided Fine-Grain Comment Extractor, which employs a gate fusion mechanism to utilize keywords generated by the Keyword Extractor, guiding the generation of code comments. Furthermore, a novel differentiation-aware enhancing encoder, which utilizes contrastive learning, is proposed to bolster the effectiveness and robustness of the KFCC model.

Process of the KFCC model and the process of data enhancement and keyword extraction

This research have been published in the prestigious journal Expert Systems with Applications, recognized by the Chinese Academy of Sciences and SCI with an impact factor of 8.5. The paper lists Zhang Rui of the Ministry of Management and Information Technology as the first author, and Zhang Chenghao as the third author. The research was supervised by Professor Yu Jianjun.

The findings of this research have been published in the prestigious journal Expert Systems with Applications, recognized by the Chinese Academy of Sciences and SCI with an impact factor of 8.5. The paper lists Zhang Rui of the Ministry of Management and Information Technology as the first author, and Zhang Chenghao as the third author. The research was supervised by Professor Yu Jianjun.