THE GEOMETRICAL LEARNING OF BINARY NEURAL NETWORKS

被引：68

作者：

KIM, JH ^{[1
]}

PARK, SK ^{[1
]}

机构：

[1] TENNESSEE TECHNOL UNIV,DEPT ELECT ENGN,COOKEVILLE,TN 38505

来源：

IEEE TRANSACTIONS ON NEURAL NETWORKS | 1995年 / 6卷 / 01期

基金：

美国国家科学基金会;

关键词：

D O I：

10.1109/72.363432

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this paper, the learning algorithm called expand-and-truncate learning (ETL) is proposed to train multilayer binary neural networks (BNN) with guaranteed convergence for any binary-to-binary mapping. The most significant contribution of this paper is the development of learning algorithm for three-layer BNN which guarantees the convergence, automatically determining a required number of neurons in the hidden layer. Furthermore, the learning speed of the proposed ETL algorithm is much faster than that of back propagation learning algorithm in a binary field. Neurons in the proposed BNN employ a hard-limiter activation function, with only integer weights and integer thresholds. Therefore, this will greatly facilitate actual hardware implementation of the proposed BNN using currently available digital VLSI technology.

引用

页码：237 / 247

页数：11