Page 96 - Anatomy of a Robot

03_200256_CH03/Bergren 4/17/03 12:27 PM Page 81
1.0
1.0
1.0 Sensory Inputs 1.0 1.0 0.8 COMPUTER HARDWARE 81
1.0
0.8
0.8 0.8 0.8 0.8
0.6 0.6 0.6 0.6 0.6 0.6
0.4 0.4 0.4 0.4 0.4 0.4
0.2 0.2 0.2 0.2 0.2 0.2
Neuron
1.0 1.0 1.0 1.0 1.0 1.0 Layers
0.8 0.8 0.8 0.8 0.8 0.8
0.6 0.6 0.6 0.6 0.6 0.6
0.4 0.4 0.4
0.4 0.4 0.4
0.2 0.2 0.2 0.2 0.2 0.2
1.0 1.0 1.0 1.0 1.0 1.0
0.8 0.8 0.8
0.8 0.8 0.8
0.6 0.6 0.6 0.6 0.6 0.6
0.4 0.4 0.4
0.4 0.4 0.4
0.2 0.2 0.2 0.2 0.2 0.2
1.0 1.0
1.0 1.0 1.0 1.0
0.8 0.8 0.8 0.8 0.8 0.8
0.6 0.6 0.6 0.6 0.6 0.6
0.4 0.4 0.4 0.4 0.4 0.4
0.2 0.2 0.2 0.2 0.2 0.2
Output Results
Judgement Score
FIGURE 3-4 One model of a neural network computer
software in systems that must develop “judgment.” One application has been the pre-
diction of credit card fraud. By exposing the neural network software to many credit card
applications and then telling the network which customers defaulted later, the network
is trained to scan new applications and reject those customers who might default later.
Here are some URLs for further study about neural networks:
www.emsl.pnl.gov:2080/proj/neuron/neural/what.html
http://vv.carleton.ca/ neil/neural/neuron.html
www.cs.stir.ac.uk/ lss/NNIntro/InvSlides.html
http://hem.hj.se/ de96klda/NeuralNetworks.htm

Special-Purpose Processors
The primary advantage of a computer is its blinding speed. It can execute many millions
of instructions every second. But some tasks require the processing of truly massive
amounts of information. These applications require the addition of even higher speed
hardware to process the information. Such high-speed hardware is specifically designed
to process the information at hand, but it can perform no other function. The high-speed