<s>
A	O
physical	B-Algorithm
neural	I-Algorithm
network	I-Algorithm
is	O
a	O
type	O
of	O
artificial	B-Architecture
neural	I-Architecture
network	I-Architecture
in	O
which	O
an	O
electrically	O
adjustable	O
material	O
is	O
used	O
to	O
emulate	O
the	O
function	O
of	O
a	O
neural	O
synapse	B-Application
or	O
a	O
higher-order	O
(	O
dendritic	O
)	O
neuron	O
model	O
.	O
</s>
<s>
"	O
Physical	O
"	O
neural	B-Architecture
network	I-Architecture
is	O
used	O
to	O
emphasize	O
the	O
reliance	O
on	O
physical	O
hardware	O
used	O
to	O
emulate	O
neurons	O
as	O
opposed	O
to	O
software-based	O
approaches	O
.	O
</s>
<s>
More	O
generally	O
the	O
term	O
is	O
applicable	O
to	O
other	O
artificial	B-Architecture
neural	I-Architecture
networks	I-Architecture
in	O
which	O
a	O
memristor	O
or	O
other	O
electrically	O
adjustable	O
resistance	O
material	O
is	O
used	O
to	O
emulate	O
a	O
neural	O
synapse	B-Application
.	O
</s>
<s>
In	O
the	O
1960s	O
Bernard	O
Widrow	O
and	O
Ted	O
Hoff	O
developed	O
ADALINE	B-Algorithm
(	O
Adaptive	B-Algorithm
Linear	I-Algorithm
Neuron	I-Algorithm
)	O
which	O
used	O
electrochemical	O
cells	O
called	O
memistors	B-General_Concept
(	O
memory	O
resistors	O
)	O
to	O
emulate	O
synapses	B-Application
of	O
an	O
artificial	O
neuron	O
.	O
</s>
<s>
The	O
memistors	B-General_Concept
were	O
implemented	O
as	O
3-terminal	O
devices	O
operating	O
based	O
on	O
the	O
reversible	O
electroplating	O
of	O
copper	O
such	O
that	O
the	O
resistance	O
between	O
two	O
of	O
the	O
terminals	O
is	O
controlled	O
by	O
the	O
integral	O
of	O
the	O
current	O
applied	O
via	O
the	O
third	O
terminal	O
.	O
</s>
<s>
The	O
ADALINE	B-Algorithm
circuitry	O
was	O
briefly	O
commercialized	O
by	O
the	O
Memistor	B-General_Concept
Corporation	O
in	O
the	O
1960s	O
enabling	O
some	O
applications	O
in	O
pattern	O
recognition	O
.	O
</s>
<s>
However	O
,	O
since	O
the	O
memistors	B-General_Concept
were	O
not	O
fabricated	O
using	O
integrated	O
circuit	O
fabrication	O
techniques	O
the	O
technology	O
was	O
not	O
scalable	O
and	O
was	O
eventually	O
abandoned	O
as	O
solid-state	O
electronics	O
became	O
mature	O
.	O
</s>
<s>
In	O
1989	O
Carver	O
Mead	O
published	O
his	O
book	O
Analog	O
VLSI	O
and	O
Neural	O
Systems	O
,	O
which	O
spun	O
off	O
perhaps	O
the	O
most	O
common	O
variant	O
of	O
analog	O
neural	B-Architecture
networks	I-Architecture
.	O
</s>
<s>
Alex	O
Nugent	O
describes	O
a	O
physical	B-Algorithm
neural	I-Algorithm
network	I-Algorithm
as	O
one	O
or	O
more	O
nonlinear	O
neuron-like	O
nodes	O
used	O
to	O
sum	O
signals	O
and	O
nanoconnections	O
formed	O
from	O
nanoparticles	O
,	O
nanowires	O
,	O
or	O
nanotubes	O
which	O
determine	O
the	O
signal	O
strength	O
input	O
to	O
the	O
nodes	O
.	O
</s>
<s>
Alignment	O
or	O
self-assembly	O
of	O
the	O
nanoconnections	O
is	O
determined	O
by	O
the	O
history	O
of	O
the	O
applied	O
electric	O
field	O
performing	O
a	O
function	O
analogous	O
to	O
neural	O
synapses	B-Application
.	O
</s>
<s>
Numerous	O
applications	O
for	O
such	O
physical	B-Algorithm
neural	I-Algorithm
networks	I-Algorithm
are	O
possible	O
.	O
</s>
<s>
Another	O
example	O
of	O
a	O
physical	B-Algorithm
neural	I-Algorithm
network	I-Algorithm
is	O
taught	O
by	O
U.S.	O
Patent	O
No	O
.	O
</s>
<s>
7,039,619	O
entitled	O
"	O
Utilized	O
nanotechnology	O
apparatus	O
using	O
a	O
neural	B-Architecture
network	I-Architecture
,	O
a	O
solution	O
and	O
a	O
connection	O
gap	O
,	O
"	O
which	O
issued	O
to	O
Alex	O
Nugent	O
by	O
the	O
U.S.	O
Patent	O
&	O
Trademark	O
Office	O
on	O
May	O
2	O
,	O
2006	O
.	O
</s>
<s>
A	O
further	O
application	O
of	O
physical	B-Algorithm
neural	I-Algorithm
network	I-Algorithm
is	O
shown	O
in	O
U.S.	O
Patent	O
No	O
.	O
</s>
<s>
7,412,428	O
entitled	O
"	O
Application	O
of	O
hebbian	O
and	O
anti-hebbian	O
learning	O
to	O
nanotechnology-based	O
physical	B-Algorithm
neural	I-Algorithm
networks	I-Algorithm
,	O
"	O
which	O
issued	O
on	O
August	O
12	O
,	O
2008	O
.	O
</s>
<s>
More	O
recently	O
,	O
it	O
has	O
been	O
argued	O
that	O
also	O
complex	O
networks	O
of	O
purely	O
memristive	O
circuits	O
can	O
serve	O
as	O
neural	B-Architecture
networks	I-Architecture
.	O
</s>
<s>
In	O
2002	O
,	O
Stanford	O
Ovshinsky	O
described	O
an	O
analog	O
neural	B-Architecture
computing	I-Architecture
medium	O
in	O
which	O
phase-change	O
material	O
has	O
the	O
ability	O
to	O
cumulatively	O
respond	O
to	O
multiple	O
input	O
signals	O
.	O
</s>
<s>
DARPA	O
's	O
SyNAPSE	B-Application
project	I-Application
has	O
funded	O
IBM	O
Research	O
and	O
HP	O
Labs	O
,	O
in	O
collaboration	O
with	O
the	O
Boston	O
University	O
Department	O
of	O
Cognitive	O
and	O
Neural	O
Systems	O
(	O
CNS	O
)	O
,	O
to	O
develop	O
neuromorphic	O
architectures	O
which	O
may	O
be	O
based	O
on	O
memristive	O
systems	O
.	O
</s>
<s>
In	O
2022	O
,	O
researchers	O
reported	O
the	O
development	O
of	O
nanoscale	O
brain-inspired	O
artificial	O
synapses	B-Application
,	O
using	O
the	O
ion	O
proton	O
(	O
)	O
,	O
for	O
'	O
analog	O
deep	B-Algorithm
learning	I-Algorithm
 '	O
.	O
</s>
