A supervised learning algorithm for feedforward networks with inhibitory lateral connections

• Main Author: Alvarez, Maria P.
• Format: text
• Published: Animo Repository 1997
• Subjects: Algorithms; Neural circuitry; Feedforward control systems; Digital computer simulation; Computer networks; Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/12111
• Institution: De La Salle University

Artificial neural network models, particularly the perceptron and the backpropagation network, do not perform lateral inhibition, a function commonly performed by biological neural networks. This paper presents a supervised learning algorithm for feedforward networks with inhibitory lateral connections. The supervised learning algorithm is developed with weight update rules for both the feedforward weights and the inhibitory lateral weights. These rules are derived mathematically using the gradient descent. The supervised learning algorithm is first developed for feedforward networks with one hidden layer and then generalized for multilayered feedforward networks with r hidden layers. Results of simulations for the XOR problem, the palindrome problem and the T-C problem are presented to validate the derived supervised learning algorithm.