Learning from Noisy Data under Distributional Assumptions
Nicolò Cesa-Bianchi, Ohad Shamir and Shai Shalev-Shwartz
In: Robust Statistical Learning Workshop, NIPS 2010(2011).
We study the framework of online learning, when individual examples are
corrupted by random noise, and both examples and noise type can be cho-
sen adversarially. Previous work has shown that without knowledge of the
noise distribution, it is possible to learn using a random, potentially un-
bounded number of independent noisy copies of each example. Moreover,
it is generally impossible to learn with just one noisy copy per example. In
this paper, we explore the consequences of being given some side informa-
tion on the noise distribution. We consider several settings, and show how
one can learn linear and kernel-based predictors using just one or two noisy
views of each example, depending on the side information provided.