论文信息 - Nonseparable data models for a single-layer perceptron

Nonseparable data models for a single-layer perceptron

This paper describes two nonseparable data models that can be used to study the convergence properties of perceptron learning algorithms. A system identification formulation generates the training signal, with an input that is a zero-mean Gaussian random vector. One model is based on a two-layer perceptron configuration, while the second model has only one layer but with a multiplicative output node. The analysis in this paper focuses on Rosenblatt's training procedure, although the approach can be applied to other learning algorithms. Some examples of the performance surfaces are presented to illustrate possible convergence points of the algorithm for both nonseparable data models.

John J. Shynk | Neil J. Bershad | J. Shynk | N. Bershad

[1] John J. Shynk,et al. Performance surfaces of a single-layer perceptron , 1990, IEEE Trans. Neural Networks.

[2] Geoffrey E. Hinton,et al. Learning internal representations by error propagation , 1986 .

[3] P J Webros. BACKPROPAGATION THROUGH TIME: WHAT IT DOES AND HOW TO DO IT , 1990 .

[4] Bernard Widrow,et al. 30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[5] John J. Shynk,et al. Stationary points of a single-layer perceptron for nonseparable data models , 1993, Neural Networks.

[6] Robert F. Pawula,et al. A modified version of Price's theorem , 1967, IEEE Trans. Inf. Theory.

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

[8] John J. Shynk,et al. Convergence properties and stationary points of a perceptron learning algorithm , 1990 .

[9] John J. Shynk,et al. Statistical analysis of the single-layer backpropagation algorithm. II. MSE and classification performance , 1993, IEEE Trans. Signal Process..