Crowd-based Debugging with Extractive Summarization of Crowd Posts

PhD Thesis Proposal Defence


Title: "Crowd-based Debugging with Extractive Summarization of Crowd Posts"

by

Mr. Fuxiang CHEN


Abstract:

Debugging is hard. Even though root causes to bugs are found, fixing them is 
non-trivial and requires a significant amount of time. For example, a previous 
study has reported that the median time to fix a single bug is 200 days. Stack 
Overflow, a question and answering forum for developers, has attracted numerous 
contributions, in the order of millions (both in asking questions and providing 
answers within short period of time), since its establishment. This makes Stack 
Overflow an invaluable source of information for developers.

In this thesis, we first propose mining the QA site, Stack Overflow, to 
leverage the huge mass of crowd knowledge to help developers in debugging their 
code. Our approach finds defective code fragments by detecting code clones 
before using them to triangulate source code anomalies. The defective code 
fragments (with the crowd’s explanation - problem-cause description) are then 
coupled with the crowd’s suggested solution (with the crowd’s explanation as 
well - solution description) and reported back to developers.

Often, the problem-case and the solution for a given software bug involve a 
large amount of textual information and code snippets. It has also been 
reported that the irrelevance and redundancy of Stack Overflow answers may 
inhibit developers’ ability to retreive information from Stack Overflow 
efficiently. Thus, to aid developers in their comprehension tasks, we next 
propose providing both problem-cause and solution summaries of the Stack 
Overflow answer posts. Our technique comprises of an ensemble of two models of 
extractive summarization techniques involving detecting salient sentences by 
making use of Mutual Reinforcement Principle and leveraging a Deep Neural 
Network architecture which aims to reduce sentence redendancy as one of its 
objectives.


Date:			Friday, 13 April 2018

Time:                  	10:30am - 12:30pm

Venue:                  Room 3494
                         (lifts 25/26)

Committee Members:	Dr. Sunghun Kim (Supervisor)
  			Prof. Andrew Horner (Chairperson)
 			Prof. Shing-Chi Cheung
  			Prof. Frederick Lochovsky


**** ALL are Welcome ****