论文信息 - An Optimality Principle for Unsupervised Learning

An Optimality Principle for Unsupervised Learning

We propose an optimality principle for training an unsupervised feedforward neural network based upon maximal ability to reconstruct the input data from the network outputs. We describe an algorithm which can be used to train either linear or nonlinear networks with certain types of nonlinearity. Examples of applications to the problems of image coding, feature detection, and analysis of random-dot stereograms are presented.

Terence D. Sanger | T. Sanger

[1] D. Pollen,et al. Relationship between spatial frequency selectivity and receptive field profile of simple cells. , 1979, The Journal of physiology.

[2] E. Bienenstock,et al. Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex , 1982, The Journal of neuroscience : the official journal of the Society for Neuroscience.

[3] Erkki Oja,et al. Subspace methods of pattern recognition , 1983 .

[4] Geoffrey E. Hinton,et al. Learning representations by back-propagating errors , 1986, Nature.

[5] R. Linsker,et al. From basic network principles to neural architecture , 1986 .

[6] Richard P. Lippmann,et al. An introduction to computing with neural nets , 1987 .

[7] R. Lippmann,et al. An introduction to computing with neural nets , 1987, IEEE ASSP Magazine.

[8] Ralph Linsker,et al. Self-organization in a perceptual network , 1988, Computer.

[9] Bernard Widrow,et al. Adaptive switching circuits , 1988 .

[10] Terence D. Sanger,et al. Optimal unsupervised learning , 1988, Neural Networks.

[11] Terence D. Sanger,et al. Optimal unsupervised learning in a single-layer linear feedforward neural network , 1989, Neural Networks.

[12] Geoffrey E. Hinton. Connectionist Learning Procedures , 1989, Artif. Intell..