New A2I2 Registered Report Accepted at Conference

News / A2I2 Media / October 4, 2022

A Registered Report from A2I2 PhD candidate Tuan Dung Lai has been accepted at the 2022 International Symposium on Empirical Software Engineering and Measurement (ESEM). 

Tuan’s report, Comparative analysis of real bugs in open-source Machine Learning projects, aims to investigate existing bugs in over 4,000 open-source software projects. 

Machine Learning (ML)-based systems possess specific characteristics that differ from traditional software systems. They rely on data to make predictions, and the data is constantly changing based on the environment and will affect the result. 

“So far, research in this area has tried to categorise different classes of faults in deep learning systems, based upon interviews with practitioners and related discussions between developers,” Tuan said.

“My research will validate the result of Nargiz Humbatova’s recent work in this field, based on empirical data of open-source projects.”

Tuan’s research will also examine traditional and ML-based software systems, with an aim to establish whether ML-based issues take longer to resolve than their traditional software counterparts.

“If my research finds that particular classes of ML issues require more effort to fix, this will give other people working in software engineering for AI (SE4AI) greater context to what issues they should be focusing on,” Tuan highlighted.

Registered Reports (RRs) are relatively new to the software engineering field. This is a promising step towards higher quality research in this area.  

The results of this study will be published at a later date in the Empirical Software Engineering journal, and will form the context for further research Tuan will be conducting.

To learn more about research at A2I2, please visit: