Loading Events
  • This event has passed.

Unsupervised Ensemble Learning

September 21, 2018 @ 11:00 am - 12:00 pm

Boaz Nadler (Weizmann Institute)


Abstract: In various applications, crowdsourcing in particular, one is given the advice or predictions of several classifiers of unknown reliability, over multiple questions or queries. This scenario is different from standard supervised learning where classifier accuracy can be assessed from available labeled training or validation data, and raises several questions: given only the predictions of several classifiers of unknown accuracies, over a large set of unlabeled test data, is it possible to:
a) reliably rank them, and b) construct a meta-classifier more accurate than any
individual classifier in the ensemble?In this talk we’ll show that under various independence assumptions between classifier errors, this high dimensional data hides simple low dimensional structures. Exploiting these, we will present simple spectral methods to address the above questions, and derive new unsupervised spectral meta-learners. We’ll prove these methods are asymptotically consistent when the model assumptions hold, and present their empirical success on a variety of unsupervised learning problems.


Biography: Boaz Nadler is a professor at the Weizmann Institute of Science. He received all his degrees from Tel-Aviv University. He then spent three years as a Gibbs instructor/assistant professor at Yale University before joining the Weizmann Institute. He also spent a sabbatical year at UC Berkeley and Stanford. His research interests are in mathematical statistics; statistical machine learning; as well as in applications in optics, signal and image processing.


MIT Statistics + Data Science Center
Massachusetts Institute of Technology
77 Massachusetts Avenue
Cambridge, MA 02139-4307