more informationMore information: Watson Libraries for Analysis | Project CodeNet | Graph4Code | Quetzal RDF Store | Jikes RVM | WALA Hackathon at PLDI 2017 | JSTools 2016 | JS Tools 2017
I have been a Research Staff Member at IBM's Thomas J. Watson Research Center since 2000, and recently became a Principal Research Staff Member. I work on a range of topics, including static program analysis, software testing, the semantic web (AI) and programming technology support for machine learning. I have also worked on the Jikes Research Virtual Machine (Jikes RVM).
- I got involved recently with Project CodeNet, and I have been developing analysis using WALA that creates graphs based on the System Dependence Graph for use with Graph Neural Network (GNN) machinery for code classification problems.
- We revealed challenges with common machine learning practice that necessitate techniques like the following: always make every tensor dimension be a different size, to minimize matrix manipulation bugs; another tactic is copious comments detailing the layout of tensors. We believe that program analysis can help obviate such burdens and make code more reliable; we have started building such support using WALA: https://github.com/wala/ML
- My testing work has been focused on several areas: Web applications in the Apollo project, finding concurrency bugs using both dynamic execution and model checking, and finding security issues in Android apps (https://dl.acm.org/citation.cfm?id=2771803)
- My semantic Web work has been on scalable inference with the SHER project; recently, I have focused on representing RDF data efficiently in an RDBMS. You can see a summary of much of our Semantic Web work in keynote I gave at the Semantic Big Data workshop at SIGMOD 2017: https://www.youtube.com/watch?v=YTSgXrGnxjg&t=2h1m44s
I was educated at the University of Wisconsin-Madison as an undergraduate, and at the University of Illinois at Urbana-Champaign as a graduate student where I worked with Professor Andrew Chien on programming systems for massively-parallel machines.
For a more complete list of publications than the one shown on this page, use DBLP or the ACM Digital Library. Use Google Scholar for more complete listings of both papers and patents.