Tue 16 May 2023 11:15 - 11:30 at Meeting Room 104 - Morning session
Current approaches to automatic program repair (APR) are largely misguided. To be truly practical, APR must be available to the programmers as a tool they can use while they are developing and testing systems. This means that it has to work within an IDE while the developer is debugging. This in turn requires that APR work without a test suite, that it work without having to rerun the system, that it can correct a good fraction of common errors, and that it is fast enough to be used interactively. We are developing a prototype system, ROSE, that demonstrates that these goals can be achieved. Our initial evaluations confirm the effectiveness and utility of our approach.
Tue 16 MayDisplayed time zone: Hobart change
Tue 16 May
Displayed time zone: Hobart change
11:00 - 12:30 | |||
11:15 15mTalk | Quick Repair of Semantic Errors for Debugging APR | ||
11:30 20mTalk | An Analysis of the Automatic Bug Fixing Performance of ChatGPT APR Dominik Sobania Johannes Gutenberg University Mainz, Martin Briesch Johannes Gutenberg University Mainz, Carol Hanna University College London, Justyna Petke University College London | ||
11:50 20mTalk | An Extensive Study on Model Architecture and Program Representation in the Domain of Learning-based Automated Program Repair APR Dániel Horváth Department of Software Engineering, University of Szeged, Szeged, Hungary, Viktor Csuvik Department of Software Engineering, MTA-SZTE Research Group on Artificial Intelligence, University of Szeged, Szeged, Hungary, Tibor Gyimóthy University of Szeged, Hungary, László Vidács University of Szeged, Hungary | ||
12:10 20mTalk | Mining Fix Patterns with Context Information for Automatic Program Repair APR Phan Thi Thanh Huyen Hitachi, Ltd., Research &Development Group, Kazuya Yasuda Hitachi, Ltd., Shinji Itoh Hitachi, Ltd., Research &Development Group | ||
12:30 90mLunch | Lunch APR |