Wednesday, May 6, 2015 - 12:00pm to 1:00pm
Location:ASA Conference Room 6115 Gates & Hillman Centers
Speaker:NIKA HAGHTALAB, Ph.D. Student http://www.cs.cmu.edu/~nhaghtal/
We study the learnability of linear separators in R^d in the presence of bounded (a.k.a Massart) noise. This is a realistic generalization of the random classification noise model, where the adversary can flip each example x with probability \eta(x) \leq \eta. We provide the first polynomial time algorithm that can learn linear separators to arbitrarily small excess error in this noise model under the uniform distribution over the unit ball in R^d, for some constant value of \eta. While widely studied in the statistical learning theory community in the context of getting faster convergence rates, computationally efficient algorithms in this model had remained elusive. Our work provides the first evidence that one can indeed design algorithms achieving arbitrarily small excess error in polynomial time under this realistic noise model and thus opens up a new and exciting line of research.
We additionally provide lower bounds showing that popular algorithms such as hinge loss minimization cannot lead to arbitrarily small excess error under Massart noise, even under the uniform distribution. Our work instead, makes use of a margin based technique developed in the context of active learning. As a result, our algorithm is also an active learning algorithm with label complexity that is only logarithmic in the desired excess error.
Joint work with Pranjal Awasthi, Nina Balcan, and Ruth Urner (to appear in COLT 2015).