
Machine Learning II (2007/2008)5/16 Computational Learning TheoryIn this overview lecture of computational learning theory we discuss such classical topics as learning boolean functions, decision lists and half planes in ndimensional Euclidean space. We describe the PAC model and the online learning model and end with an introduction to VC theory. Concepts covered: Nonprobabilistic approaches to learning, online learning and PAC learning. Learning conjunctions, disjunctions and decision lists. The eliminiation algorithm, winnow, the weighted majority algorithm and the perceptron. Growth function and VapnikChervonenkis dimension.
Slides available [here] 