Credentials: Ph.D. Student, Electrical and Computer Engineering
LUCID Research Project: NEXT
Research Interests: large scale machine learning, distributed optimization, active learning
Research Website: Scott Sievert
Research Advisor: Rob Nowak
Hello. I’m a Scott Sievert, a graduate student here at Madison, and I work on LUCID with optimization and machine learning.
Specifically, I help with NEXT, a system to apply active learning to crowdsourcing. This can bring very large benefits, mostly that of reduced costs to run psychology experiments, for example.
Perhaps the best part is that psychology researchers have used this and have generated novel research questions for us. So that it is mutually beneficial. We’ve made their work more efficient and they have given us questions to ask.
This work on NEXT has inspired other interests, now I’m interested in the systems side of machine learning.
That is, what modifications to the theory can I make to perform and observe some practical benefit?
One example of this is that training a machine learning model takes a long time. A research project of mine this semester was trying to minimize the time it takes to train using some small, but important theory changes.