We determine the complexity of learning problems for unary regular languages. We begin by investigating the minimum consistent dfa (resp. nfa) problem which is known not to be app...
Statistical learning theory chiefly studies restricted hypothesis classes, particularly those with finite Vapnik-Chervonenkis (VC) dimension. The fundamental quantity of interest i...
We present a simple, agnostic active learning algorithm that works for any hypothesis class of bounded VC dimension, and any data distribution. Our algorithm extends a scheme of C...
The paper consists of two major parts. In the first part, we re-examine relative -approximations, previously studied in [12, 13, 18, 25], and their relation to certain geometric p...
We study model selection strategies based on penalized empirical loss minimization. We point out a tight relationship between error estimation and data-based complexity penalizatio...