Hi. I'm Chaiyong Ragkhitwetsagul.
I'm currently an (almost) 3rd year MPhil/PhD student in Centre for Research on Evolution, Search and Testing (CREST) under Software Systems Engineering (SSE) group in the Department of Computer Science at University College London. I'm also an assistant instructor at the Faculty of ICT, Mahidol University, my sponsor, which I'll go back to continue working after completing my PhD.
I received a bachelor degree in Computer Engineering from Kasetsart University, Thailand, and a master degree of MSIT-Very Large Information Systems from Carnegie Mellon University, USA.
I'm currently staying in London, a wonderful city except the rain (especially in winter) ...
I'm generally interested in software engineering research but specifically on these three topics:
Source Code Similarity
My PhD research topic is "Measuring Code Similarity in Large-scaled Code Corpora". I am currently finding a technique to retrieve similar source code from over web-scale datasets (something at the scale of over 100M SLOC) with high accuracy. I'm also working on an empirical analysis to compare various state-of-the-arts source code similarity analysers against different types of code modifications.
I am particularly interested in clone detection and its related work or applications. I've worked on using search-based techniques with clones such as optimising clone detection tool's parameters to maximise clone agreemenet among tools
Mining of Software Repositories
I am currenlty studying the effects of cloned code between Stack Overflow and open source projects in terms of licensing violation and outdated code.
- 10 Jan 2017: "Using Compilation/Decompilation to Enhance Clone Detection" is accepted at IWSC 2017. We compile and decompile three software systems and found that the comp/decompilation process helps increasing clone recall of NiCad without sacrificing precision.
- 25 Oct 2016: My SCAM'16 paper is invited for EMSE journal extension! We are excited to expand the experiment further to another dataset and see if we can reuse the optimal configurations derived from the first dataset we have in the SCAM paper to it.
- 12 Sep 2016: I have received the NSF Student Travel Support for my trip to SSBSE 2016 in the US. Thanks to the organiser!
- 29 Jul 2016: My submission of "Similarity of Source Code in the Presence of Pervasive Modifications" to SCAM 2016 is accepted. This is my very first full research paper. It is the work I have been working on for more than a year. We have evaluated 30 similarity analysers against 50 Java source code files that have been pervasively modified using code obfuscators/compiler/decompilers. We found that specialised tools such as CCFinderX clone detector has better performance than general tools such as string similarity and NCD tools.
- 29 Jul 2016: My submission of "Measuring Code Similarity in Large-scaled Code Corpora" to late pre-doctoral ICSME 2016 Doctoral Symposium is accepeted. I'm thrilled to discuss my PhD work with other research students at ICSME.
- 30 Jun 2016: I have been awarded $20,000-worth cloud computing power from Microsoft Azure for Research for 1 year for my proposal of "ISiCS: Internet-scaled Similar Code Search". We're planning to use it for a large-scale experiment. It is going to be fun!
- 23 Jun 2016: My SSBSE Challenge Track 2016 paper "Searching for Configurations in Clone Evaluation: A Replication Study" co-authored with Matheus Paixao, Manal Adham, Saheed Busari, Jens Krinke, and John H. Drake is accepeted. I was very excited since this is the first research paper ever in my PhD study. We found that EvaClone, clone detector's parameter optimisation framework, can give higher agreements among six tools but with too many clones reports. The fitness function needs improvements.
You can find my CV here:
First: Dr. Jens Krinke
Second: Dr. David Clark