NTT Software Innovation Center Collaborates with Prof. André van der Hoek on Recovering Undocumented Knowledge

Dr. Shinobu Saito, Prof. Andre van der Hoek, undergraduate researcher Taneisha Kaur Arora, Ph.D. student Adriana Meza Soria, and Project Manager Mr. Daisuke Hamuro of NTT SIC.A common problem in both open source and closed source software development is the checking in of code changes that are not linked to any issues in the issue tracker. Especially when an issue does exist that actually describes the needed change in functionality, such missing links represent knowledge that should be captured but is not. NTT is no stranger to this problem and recognizes the importance of addressing it.

For the past year, Dr. Shinobu Saito from the NTT Software Innovation Center has been working with Prof. André van der Hoek and his team to do exactly that. Key to the project has been Taneisha Kaur Arora, an undergraduate majoring in both Software Engineering and Data Science, who worked together with van der Hoek’s graduate student Adriana Meza Soria to develop a tool suite to perform experiments on an open source data set exhibiting the problem. Using her outstanding knowledge of machine learning, information retrieval, and statistics, Arora experimented with a range of different approaches, slowly but steadily making progress in more robustly recovering the links. In their most recent experiments, the team achieved a success rate above 80% with several refinements still to be made. Naturally, a next step is applying the approach to an NTT data set, which Dr. Saito and his colleagues at NTT SIC in Japan are now preparing.

“It has been tremendously rewarding collaborating with Dr. Saito for the past few years,” van der Hoek says. “The ability to work on problems that are directly relevant to the real world is of great value, and brings to our students a challenge they not only embrace, but also can turn into their next job!” Arora, indeed, is off to Google this summer for an internship that in no small measure emerged because of her experience working on this problem with NTT.

This fruitful collaboration with Dr. Saito, a Distinguished Research Engineer, Software Engineering at NTT SIC, grew out of Saito’s time spent at ISR as a visiting researcher from July 2016 – June 2018.

For more information, contact Prof. van der Hoek.

This article appeared in ISR Connector issue: