论文信息 - High Capacity Neural Networks for Familiarity Discrimination

High Capacity Neural Networks for Familiarity Discrimination

This paper presents two new novelty discrimination models for uncorrelated patterns based on neural modelling. The first model uses a single neuron with Hebbian learning and works well when the number of memorised patterns is less than 0.046N (N, the number of inputs). The second model is based on checking the value of the energy function of a Hopfield network. By sacrificing the ability to extract patterns - not needed for familiarity detection - an amazing jump from the normal capacity for retrieval of 0.145N to a capacity for novelty discrimination of 0.023N2 is achieved. In addition, both models give some insight on the effect of deja vu, since there is always a very small probability of detecting novel patterns as familiar.

[1] J J Hopfield,et al. Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[2] R. Palmer,et al. Introduction to the theory of neural computation , 1994, The advanced book program.

[3] Stephen Grossberg,et al. Familiarity Discrimination of Radar Pulses , 1998, NIPS.

[4] Gail A. Carpenter,et al. Threshold Determination for ARTMAP-FD Familiarity Discrimination , 1997 .

[5] Lucas C. Parra,et al. Statistical Independence and Novelty Detection with Information Preserving Nonlinear Maps , 1996, Neural Computation.

[6] Robert P. W. Duin,et al. Novelty Detection Using Self-Organizing Maps , 1997, ICONIP.

[7] John J. Hopfield,et al. Neural networks and physical systems with emergent collective computational abilities , 1999 .

[8] R. Palmer,et al. , Introduction to the Theory of Neural Computation 1 , 2007 .