A comparison of two learning philosophies

The authors first present a learning model using class 2 dynamical systems and a learning model using class 3 dynamical systems. Then they compare these two approaches, and emphasize a unification of the two theories. The similarities of these approaches are that both schemes use the idea of storing information in stable configurations of dynamical systems. Both schemes follow the procedures of learning as encoding, change, and quantization. The algorithms used in both approaches are similar. The differences are that one model classifies input stimulus by its corresponding attractors while the other classifies input stimulus by quantization in internal parameter space. One model uses a distance defined as an inference guidance while the other uses the Hausdorff distance as an inference guidance