Neural Networks
NTT Systems provided research assistance from 1989-1995 to Dr. Martin Taylor of
DCIEM on numerous neural network
projects. "Neural networks" is a broad term covering many techniques in Artificial
Intelligence and Cognitive Science that attempt to model human thinking in a fashion
inspired by the architecture of the human brain itself--a vast web of interconnected
nerve cells.
NTT Systems assisted Dr. Martin Taylor in investigating the use of "chaotic" neural
networks to efficiently store memories (perhaps in a manner much like the human brain).
We also did some preliminary research into "mating" the individual neurons with each
other, treating the entire neural network like a gene pool, in an attempt to get them
to grow and adapt.
Neural control networks are a special kind of neural network based on Perceptual
Control Theory, as pioneered by W.T. Powers. We built
tools for doing research on control networks, and have
performed preliminary experiments at getting such networks
to track a simplified "toy" speech system.
Kohonen networks "discover" interesting features of the data set. NTT Systems built a
Kohonen network for use in a speech recognition system for Dr. Martin Taylor.
NTT Systems conducted exploratory research into the Finch-Chater algorithm for
classifying words into semantic categories. The algorithm is under consideration for
use in future hypertext systems (for instance, intelligent Web searching).
Please visit our home page:
NTT Systems Inc.