The use of hints as an aid in learning from examples is addressed. Hints describe the situation where, in addition to the set of examples of some unknown function f that we are trying to learn, we have prior knowledge of certain facts about f. The use of hints, under different names is coming to the surface in a number of areas dealing with learning and adaptive systems. The most common complaint is that hints are heterogeneous and cannot easily be integrated into learning. The final report describes the development of a systematic method that integrates different types of hints in the same learning process. Algorithms for learning from hints are presented. These algorithms use fixed or adaptive schedules to determine the turn of each hint to be learned in order to achieve balance among the errors of different hints. Also, a theoretical analysis of learning from hints is developed. It is based on the Vapnik-Chervonekis VC dimension, which is an established tool for analyzing learning from examples.