We present a method for learning classification functions From pre-cla
ssified training examples and hypotheses written roughly by experts. T
he goal is to produce a classification function that has higher accura
cy than either the expert's hypotheses or the classification function
inductively learned from the training examples alone. The key idea in
our proposed approach is to let the expert's hypotheses influence the
process of learning inductively from the training examples. Experiment
al results are presented demonstrating the power of our approach in a
variety of domains.