Praise for Trading from Your Gut


S in the Cambria typeface, an  S


Download 1.25 Mb.
Pdf ko'rish
bet15/82
Sana14.11.2023
Hajmi1.25 Mb.
#1772599
1   ...   11   12   13   14   15   16   17   18   ...   82
Bog'liq
Curtis Faith Trading from Your G

in
the Cambria typeface, an 
in the Comic Sans MS Bold typeface,
and an in the Sante Fe LET typeface. For the neural network to
function properly, the model contained in its neural connections
needs to capture the essence of the letter in an abstract sense.
Consider what would happen to the neural network if it was pre-
sented with the number 5. Would it be able to distinguish between
C
HAPTER
2 • T
HE
P
URPOSE OF
G
UT
I
NTUITION
25
From the Library of Daniel Johnson


ptg
the letter and the number 5? It depends on the sophistication of
the model. If the model incorporated the concept that the number 5
has sharp angles in the top with soft curves on the bottom, and the
letter has soft curves on both top and bottom, then the model
would incorporate enough knowledge to be able to properly distin-
guish between them. 
Training a neural network to properly distinguish between let-
ters and 5 numbers requires examples that incorporate the charac-
teristics that distinguish between the two symbols. If a neural
network has been trained with examples of only letters and no num-
bers, then it will not be able to distinguish between and 5. The
choice of examples must be sufficiently broad for the training to be
effective.
The choice of what examples to feed to a neural network and
the ranking for how well each of the examples fits with the model is
a left-brain activity. The left brain also analyzes the complexity of a
particular problem and makes assessments of potential models. For
example, the left brain would analyze the similarities between and
5 and realize that at least one extra layer of complexity in the model
is required to distinguish between these similar characters. The
neural network’s pattern recognition is a right-brain activity. After
using examples to configure and train a neural network, it can auto-
matically recognize patterns and perform categorization. In a com-
puter or in the human brain, it can do this almost instantaneously.
The left brain is good at extracting abstract models from pat-
terns and examples, and at establishing and choosing categories.
The right brain is skilled at determining whether a given example
fits the model that defines a category and is good at recognition. The
left brain analyzes and the right brain notices.
26
T
RADING FROM
Y
OUR
G
UT
From the Library of Daniel Johnson


ptg
The left brain analyzes and the right brain notices.

Download 1.25 Mb.

Do'stlaringiz bilan baham:
1   ...   11   12   13   14   15   16   17   18   ...   82




Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling