I have decided to change my research focus. I am now going to pursue Machine Learning.
This has been coming for a while now. I have not been truly excited about a programming language research project in a long long time. While much of research is drudgery, you do have to have some moments of inspiration to keep going. Those simply haven't been happening.
On Monday I sat down and took stock of my situation. What is it that I really like to do with computer science? Given some free time (or just when procrastinating), what do I do? I have long wished that my field required the use of statistics: I love using data to try to tease out correlations and probabilities. I also like experiments, databases, and I am absolutely fascinated by genetic and evolutionary computing---it's like playing with God's toolbox. From languages I have learned about types, functions, representations, and models. I've been programming databases, playing with genetic algorithms, and studying statistics on the side. Put all these things together, and what emerges is machine learning: it covers all of these things.
On Wednesday I went to visit a good friend Bill Hsu, who has been in this field for quite some time. I did not realize when we made our appointment that I would be making this decision, but it was certainly a nice coinciding. He gave me a quick introduction, took me to the closing ceremony of a machine learning summer school his students were attending, and gave me a list of books to read to get started.
One great thing about being an academic is that by saying the word, I can change my job description. Instead of feeling like I'm procrastinating when I write a simulation, now all of a sudden, I'm being productive!
So where to go from here. I'm now reading Machine Learning, by Tom Mitchell, and have three more books in queue. I estimate it will take a number of months to read them and start to get up to speed. After (or during, more likely) reading those, I will be able to start reading current publication, playing with my own experiments and simulations, and seeing where the open problems are. The fact that my school has quite a few faculty already in this area should be a great asset.
This is the first time in a very long time I've felt excited about doing research.