My research focuses on the correctness of software. The technique that I'm looking at is called deductive verification which requires the user to write specifications for a program. This requires a lot of time and knowledge and is therefore seen as a bottleneck in the verification process. I am trying to minimise this burden by looking into specification generation to automatically generate the specifications, so that the users don't need to write them.
I am also very interested in the application of formal methods to software security.