Explores the gap between empirical research on human learning and models from machine learning. Covers symbolic, statistical, semantic space, neural network, unsupervised, and other approaches.