I'm a computer scientist with expertise in software engineering, human-computer interaction, and machine intelligence -- "Smart Useful Systems".
I have a PhD from the University of Cambridge, and I have been a senior engineer with Australia's centre of excellence in ICT research (NICTA) and also in industry. I am also an adjunct lecturer in the University of Queensland, where I taught project management to the Advanced IT Project students in 2010.
These are the most recent projects I've been developing or working on...
What if you could explore what happened in a conversation without even listening to it? What if you could turn communication skills from an art into a science? I led the development of software that can automatically analyse and visualise aspects of a conversation from a video and transcript. This is a meaningful application of machine analysis to a human domain, and cuts the time it takes someone to examine or explore a conversation dramatically.
Applications include medical training, call centre quality analysis, social psychology and language research.
Textbooks are expensive to produce and buy, and usually don't exactly fit the course you are studying. What if we had textbooks that quicker to produce, smarter, work better with the way students learn and teachers teach, and that can make use of the whole world of information out there?
Shortlisted for the 2010 Telstra Innovation Challenge
How do you get novice students learning difficult mathematics to be able to relate to a machine reasoning system that is so complex that usually only PhD researchers go near it? How can you get two groups that don't understand how each other thinks at all (people and machine reasoning systems) to co-operate effectively on difficult problems?
Multiple choice questions give the answer away to the student. Short answer questions are hard for computers to mark (complex natural language analysis). What if there was a simple kind of question that meant the student had to come up with the answer, but that a computer could mark automatically with no ambiguity or inaccuracy?
US Patent Application