The 33rd IEEE/ACM International Conference on Automated Software Engineering (ASE 2018) was held in Montpellier, France from September 3rd to September 7th. A total of three papers by the College of Computer Science and Technology of Zhejiang University were accepted. Among them, the paper Neural-Machine-Translation-Based Commit Message Generation: How Far Are We? won the ACM SIGSOFT Distinguished Paper Award. A total of 386 papers were submitted for this conference. After rigorous screening, 69 were accepted, and only 6 were selected as Distinguished Papers.
The award-winning paper examines the current overuse of deep learning in software engineering field. It probes into recent work that used neural machine translation algorithm to automatically generate code submission logs. It was found that the superior performance of this work was mainly attributed to the noise data in its data set. The paper proposes a simple method based on nearest neighbor algorithm called NNGen to automatically generate code submission logs. In comparison to the complex and time-consuming neural machine translation algorithms, NNGen is simpler, faster, and has significantly better performance. Through this analysis, the paper suggests that researchers should consider using simple methods to understand data and design algorithms before applying deep learning to perform software engineering tasks, avoiding overuse of deep learning algorithms.
The lead author of this award-winning paper is LIU Zhongxin, a master’s student at the College of Computer Science and Technology. The main co-authors include Dr. XIA Xin from Monash University in Australia and Prof. WANG Xinyu from the College of Computer Science and Technology at Zhejiang University.
The software engineering research team of ZJU's College of Computer Science and Technology has rich research experience in software engineering theories. It has delivered world-class research in software analytics, software warehouse mining and empirical software engineering. In recent years, the team has published more than 20 papers in CCF A journals and conferences in the field of software engineering. This is the second time the team won ACM SIGSOFT Distinguished Paper Award after receiving it at the ICSE2018 conference earlier this year.