Simultaneous Learning and Covering with Adversarial Noise
Andrew Guillory, Jeff Bilmes · 2011
We study simultaneous learning and covering problems which are set cover problems that depend on the solution to a learning problem. The goal is to jointly minimize the cost of both learning and covering. We extend previous work to allow for a limited amount of adversarial noise. As a special case our problem can be applied to noisy query learning. 1