Well, it's finally happened. I'm am officially a postgraduate student at the University of Melbourne. Here's the (currently very useless) website and email addresses [r.kelly(at)pgrad.unimelb.edu.au, rfk(at)cs.mu.oz.au] to prove it. I will be researching "Reasoning in Agent Programming Languages" under Dr Adrian Pearce. Read on if you want to know what that actually means...
Agent programming languages allow us to construct programs in terms of autonomous 'agents' using high-level concepts such as beliefs, goals, and desires. They are naturally suited to concurrent, highly-connected, unpredicatable environments (like, say, the Internet) and, like all new language technologies, are expected to make programs more powerful, more reliable and quicker to develop. Check out the Melbourne Uni Agentlab website for more details.
These programming languages are supported by a strong logical "Theory of Agency" which allows reasoning at two levels - agent programs can reason about themselves and their environment to determine how to behave, and we can reason about the behavior of our programs. It should be possible to make statements such as "X will eventually occur" or "Y will never be the case" given the specification of an agent program in an appropriate language.
My project, while still being defined, will be investigating ways to allow more powerful reasoning on both these levels.