Abstract:

Program debugging is a procedure to locate faults in programs, repair them,
and confirm the modified programs having removed the located faults. Fault
localization is a precursor of fault repair and subsequent validations. To
leverage the computation power of the underlying hardware platforms, in the
recent decade, researchers propose techniques that employ statistical
approaches to tackling the fault localization problem. These techniques
typically find the fault-related positions in programs by comparing among
the execution statistics of passed executions and failed executions to
assess the fault suspiciousness of individual program entities separately.
They ignore how infected program states may propagate via the executions of
faulty programs. In this talk, I will present our recent research efforts on
statistical fault localization. I will firstly present a technique that uses
edge profiles to represent passed executions and failed executions,
contrasts these profiles to model how each basic block contributes to the
observed failures, and propagate these contributions among basic blocks on
program dependence graphs. Second, I will describe several issues that
affect the performance of statistical fault localization techniques,
including coincidental correctness, and assumptions of statistical models,
and present our ongoing research that tackles these issues.

Biography：

Wing-Kwong Chan is currently an assistant professor at Department of
Computer Science, City University of Hong Kong. He received his BEng in
Computer Engineering, MPhil and PhD degrees all from The University of Hong
Kong. He was in the industry to manage and develop software applications for
more than 10 years prior to returning to The University of Hong Kong to
complete his PhD study. He is currently an editorial board member of Journal
of Systems and Software, a program chair of QSIC 2010 and AST 2010, a guest
editor of international journals, and a program committee member of many
international conferences. His recent research interest is to address
software analysis and testing challenges in service-oriented software,
self-adaptive software, concurrent software, and utility programs. His work
has been reported by over 60 articles at major publication venues such as
TSE, TOSEM, CACM, JSS, IST, STVR, ICSE, FSE, ASE, WWW, and ICDCS.