We describe a novel framework for the design and analysis of online learning algorithms based on the notion of duality in constrained optimization. We cast a sub-family of universa...
A fundamental open problem in computational learning theory is whether there is an attribute efficient learning algorithm for the concept class of decision lists (Rivest, 1987; Bl...
We prove logarithmic regret bounds that depend on the loss L∗ T of the competitor rather than on the number T of time steps. In the general online convex optimization setting, o...
We define notions of stability for learning algorithms and show how to use these notions to derive generalization error bounds based on the empirical error and the leave-one-out e...
One way to quantify how dense a multidag is in long paths is to find the largest n, m such that whichever n edges are removed, there is still a path from an original input to an ...