Page 85 - Artificial Intelligence in the Age of Neural Networks and Brain Computing

3. MFE Gradient Descent 73

FIGURE 3.7
Nonlinear threshold logics of activation firing rates (A) for the output classifier, (B) hidden
layers hyperbola tangent.

.
. .
X X sin h X
. . e e .
1 a ¼ tan iX ¼ . . ¼ ¼ tan h X 1;
.
X
e þ e X cos h X
d tan hðxÞ 2
¼ 1 tan hðxÞ (3.15b)
dx
We have reviewed how to appreciate the BNN. Albert Einstein’s brain has been
kept after he passed away. He had 10 billion neurons just like we do, but he has 100B
glial cells that is important for house cleaning servant function to minimize Demen-
tia Alzheimer Disease (DAD), which might have made him different from some of
us. These housekeeping smaller glial cells surrounded each neuron output called
axon that can keep positive ion vesicles move forward in a pseudo-real time which
show repulsion to one another in line, as one ion is pushed in from one end of the
axon, so that those conducting positive charge ion vesicles have no way to escape
but line up by those insulating glial cells in their repulsive chain in about 100 Hz,
100 ions per second, no matter how long or short the axon is. The longest axon is
about 1 m longer from the neck to the toe to instantaneously issue the order from
homo sapiens to run away from the tiger. The insulated fatty acids and myelin sheath
are known to be glial cells, among those six types of glial cells.The glial cell (glue
force) is derived for the first time [1] when the internal energy E int. is expanded as
.
the Taylor series of the internal representation S i related by synaptic weight matrix
. .
[W i,j ] to the power of the pairs S i ¼ W i;j X pair of which the slope turns out to be
biological glial cells identified by Donald O. Hebb learning rule
* +
. vH int:
D E
g j ¼ . ; (3.16)
vD j
.
where the j-th dendrite tree sum D j of all i-th neurons whose firing rates are in pro-
portional to the internal degree of firing rate S i called the entropy uniformity:
* +
. X .

D j ¼ W i;j S i
i