Learning-induced oscillations in a homogeneous neural network

Author

Chang, Hung-Jen ; Ghosh, Joydeep ; Liano, Kadir

Author_Institution

Dept. of Electr. & Comput. Eng., Texas Univ., Austin, TX, USA

Volume

3

fYear

1992

fDate

7-11 Jun 1992

Firstpage

474

Abstract

A mathematical model is developed to characterize the aggregate behavior of large neural networks. The model describes the global effects of weight changes brought about by a local Hebb-type adaptation rule. In particular, such adaptation can lead to rhythmic behavior of ensemble activity even in an isolated cell assembly of homogeneous cells. Conditions that make such oscillatory behavior possible are identified, and the frequency, amplitude and center of oscillation are quantitatively related to the network parameters. Results from computer simulation of a 100-neuron assembly agree closely with the mathematical predictions

Keywords

Hebbian learning; neural nets; 100-neuron assembly; aggregate behavior; ensemble activity; homogeneous neural network; large neural networks; learning-induced oscillations; local Hebb-type adaptation rule; rhythmic behavior; Assembly systems; Biological neural networks; Biological system modeling; Biomembranes; Cells (biology); Frequency; Intelligent networks; Learning systems; Neural networks; Neurons;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1992. IJCNN., International Joint Conference on

Conference_Location

Baltimore, MD

Print_ISBN

0-7803-0559-0

Type

conf

DOI

10.1109/IJCNN.1992.227129

Filename

227129