Reconceiving Machine Learning
Machine learning is a cottage industry, not an engineering discipline.
Nearly every new problem is solved from scratch. There is no reuse of
previous solutions. There is no language to describe problems. Research is
technique-oriented rather than problem-oriented.
This lack of modularity and composability limits the field advancing.
Reinvention is rife.
My research agenda is to develop a
composable basis for machine
learning. My detailed plan is partially explained in this formal
research proposal. This is now
funded and I am looking for a postdoc to work with me, Ralf Herbrich,
Ule von Luxburg and Peter Grünwald on the
project. If you are interested in the position, I encourage you to
read the proposal, and my recent papers to get a feel for the
sort of thing I am doing. The positions will be formally advertised very soon
(early September 2010)
As the proposal says, I want to focus on problems rather than solutions. I
recently came across a great quote that sums up my feeling on this:
The formulation of
a problem is often
more essential than its solution, which may be merely a matter of
mathematical or experimental skill. To raise new questions, new
possibilities, to regard old problems from a new angle, requires
creative imagination and marks real advance in science.
Progress to Date
In order to get a feel for the sort of things done so far, look at
these recent publications.