winner-take-all neural networks (wtanns)

These networks are useful for unsupervised learning tasks and applications requiring competitive learning.