A Survey on Deep-Learning-based Test Oracle Generation

PhD Qualifying Examination


Title: "A Survey on Deep-Learning-based Test Oracle Generation"

by

Mr. Tsz On LI


Abstract:

Testing is an essential task in software engineering. Performing testing 
requires generating "test oracle": a component of a test case that determines 
whether a software's behavior meets expectation; unexpected behavior indicates 
the presence of bugs. However, automatically inferring a software's expected 
behavior for an arbitrary input is often an undecidable problem. Hence, test 
oracle generation remains an outstanding challenge in automated software 
testing.

With recent advancement of deep learning, more attention has been paid to 
deep-learning-based test oracle generation. This survey conducts a systematic 
literature review on twenty-third recent papers along this research direction. 
It divides these papers into three categories, namely training-based 
approaches, inference-based approaches and empirical studies of evaluation 
methodology. Each category addresses a unique research challenge. We then 
describe the idea and contribution made by papers in each category. Finally, we 
highlight two open research challenges and opportunities.


Date:                   Friday, 30 August 2024

Time:                   3:00pm - 5:00pm

Venue:                  Room 3494
                        Lifts 25/26

Committee Members:      Prof. Shing-Chi Cheung (Supervisor)
                        Dr. Shuai Wang (Chairperson)
                        Dr. Dongdong She
                        Dr. Jiasi Shen