Seminar by Dr. Jinqiu Yang (University of Waterloo)

Speaker: Dr. Jinqiu Yang (University of Waterloo)                                                                                    

Title:  Utilizing Software Data for Automated Bug Fixing

Date: Thursday May 24, 2018

Time: 10:30 am - 12 noon

Room: EV 2.260

ABSTRACT

Software bugs significantly impair software reliability. In 2016, the estimated cost of software failures was around 1.1 trillion US dollars. Due to the large number of bugs that are reported every day, developers have limited time and resources to fix every single bug. As a result, bugs take months to be fixed on average and many remain unfixed. Automated bug fixing (also commonly referred to as automated program repair) is proposed to automatically generate patches at the source-code level. Automated bug fixing aims to free developers from the tedious and labour-intensive bug fixing activities so that developers can focus on more creative tasks.

In this talk, I will introduce my work on advancing automated bug fixing research by utilizing software data, particularly bug-fixing history. The fact that bug fixes are repetitive opens the door to leverage past bug fixes for fixing new bugs. However, there are challenges in how to accurately utilizing bug fix patterns. I will focus on two work that leverage text analytics and code analysis to address the challenges accordingly: 1) using manually-defined bug fix patterns to generate fixes from bug reports which are in natural language; and 2) using automatically-learnt bug fix patterns to generate recurring fixes. The two approaches are applied to fix bugs in large and mature real-world software systems. The evaluation shows that utilizing past bug fixes has a significant advantage compared to the state-of-the-art repair techniques in generating more high-quality fixes. I will also discuss the challenges and promising directions for future research.

BIO

Jinqiu Yang received her PhD (2018) in University of Waterloo. She received M.Sc. (2013) from the University of Waterloo and B.Eng (2011) from Nanjing University in China. Her research interests include automated program repair, improving the usability of static bug detection techniques, mining software repository and software engineering; with a focus on using software data, such as past bug fixes, to advance automated bug fixing to better assist developers in improving software quality. Her work has been published in premier venues such as FSE, ASE, EMSE, ICST and MSR. She has been collaborating with IBM since 2014 and interned as a research assistant at IBM T.J. Watson Research Center in 2016. Her work, which aims to improve the leading commercial software security product (AppScan Source), is recognized by ‘IBM CAS Project of the Year’ award in 2015.