Sandor Dornbush
computer science phd student
My areas of study include:
  • ubiquitous computing
  • machine learning
  • data mining
  • distributed systems

I have chosen these areas of research because each field has a large number of unsolved problems which will change our lives in the near future. As a professional developer I learned that at the wrong company coding can be a lot like plumbing. You connect together enough pipes and data will flow from point a to b.

With data mining and machine learning it is rarely that simple. Often the path is not known and exploration and experimentation is needed to construct a model that is effective and computationally efficient.

In my view keyboards and mice are very crude tools for interacting with computers. As we move forward many applications will be built on non-traditional Human-Computer-Interaction. In these applications computers are much more aware of their surroundings and can provide much richer interactions. In such pervasive computing environments where sensor errors are common, we must use statistical methods such as machine learning to overcome these errors.

In the world that we live we are producing huge amounts of data. Much of that data is discarded. By carefully analyzing that discarded data many interesting trends can be discovered. It is my goal to use those trends to make useful systems. Here are a couple of systems:
  • A music player that watches what you listen to and customizes your playlist according to patterns it finds in your listening behavior.
  • A GPS navigation device that collects driver speed to calculate traffic.
  • A mobile phone that monitors the ringer status (on, vibrate, off) and automatically changes the status of the ringer.