6.034 Artificial Intelligence - Recitations, fall 2004 online slides on learning
	Neural Network Issues Usefulness Generalizability Understandability (cannot explain results) Self-structuring neural networks Add/delete nodes and links until error is minimized Networks and expert systems Can we learn rules corresponding to network? Input background knowledge Predefined structure, weights Computational complexity Blum and Rivest in 1992 proved that training even a three-node network is NP Complete Inductive bias is smooth interpolation between data points