An invitation to computational learning theory

In the following period I will be overviewing some of the ideas and results from the area of Computational Learning Theory.
This field of research deals with formalizing the main aspects of the theory of statistical classification and machine learning. We are especially interested in formalizing the three key aspects : a) definition of successfully completed learning task , b) sample complexity and c) time complexity of the learning algorithm.
Apropriate theoretical framework can provide us with the fundamental results on what can be learned efficiently (in polynomial time) and how to construct the adequate learning algorithms.