Neuromorphic Chips Black Holes:

Neuromorphic Chips  Black Holes:
X-RAYS FROM LIGHTNING • DATA MINING FOR GENETIC TREASURE
Do-It-Yourself
Black Holes:
Physics Gets Ready
MAY 2005
W W W.SCIAM.COM
Neuromorphic
Chips
Differences in
Male and Female
Brains
The Weird
Warmth of
Asteroids
Stopping an
Invisible Epidemic
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
N
56
e
o
r
u
SCIENTIFIC A MERIC A N
o
m
c
r
i
o
M
c
h
c
i
i
p
h
s
p
r
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
M AY 2 0 0 5
Compact, efficient electronics based on the
brain’s neural system could yield implantable
silicon retinas to restore vision, as well
as robotic eyes and other smart sensors
n n n n
BY KWABENA BOAHEN
hen IBM’s Deep Blue supercomputer edged out
world chess champion Garry Kasparov during
their celebrated match in 1997, it did so by
means of sheer brute force. The machine evaluated some 200 million potential board moves a second,
whereas its flesh-and-blood opponent considered only three
each second, at most. But despite Deep Blue’s victory, computers are no real competition for the human brain in areas
such as vision, hearing, pattern recognition, and learning.
Computers, for instance, cannot match our ability to recognize a friend from a distance merely by the way he walks.
And when it comes to operational efficiency, there is no contest at all. A typical room-size supercomputer weighs roughly 1,000 times more, occupies 10,000 times more space and
consumes a millionfold more power than does the cantaloupe-size lump of neural tissue that makes up the brain.
W
CREDIT
How does the brain— which transmits chemical signals between neurons in a relatively sluggish thousandth of a second— end up performing
some tasks faster and more efficiently than the most powerful digital
processors? The secret appears to reside in how the brain organizes its
slow-acting electrical components.
The brain does not execute coded instructions; instead it activates
links, or synapses, between neurons. Each such activation is equivalent
to executing a digital instruction, so one can compare how many connections a brain activates every second with the number of instructions a
computer executes during the same time. Synaptic activity is staggering:
10 quadrillion (1016) neural connections a second. It would take a million
Intel Pentium-powered computers to match that rate— plus a few hundred
megawatts to juice them up.
Now a small but innovative community of engineers is making significant progress in copying neuronal organization and function. Researchers speak of having “morphed” the structure of neural connections
into silicon circuits, creating neuromorphic microchips. If successful, this
work could lead to implantable silicon retinas for the blind and sound
processors for the deaf that last for 30 years on a single nine-volt battery
IMPL ANTABLE SILICON RETINA, shown in this artist’s conception, could emulate the
eye’s natural function, restoring vision for patients with certain types of blindness.
w w w. s c ia m . c o m
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
SCIENTIFIC A MERIC A N
57
the need for the brain to expend a great deal of effort. We
chose the retina because that sensory system has been well
documented by anatomists. We then progressed to morphing
the developmental machinery that builds these biological circuits — a process we call metamorphing.
SILICON RETINA senses the side-to-side head movements of University
of Pennsylvania researcher Kareem Zaghloul. The four types of silicon
ganglion cells on his Visio1 chip emulate real retinal cells’ ability to
preprocess visual information without huge amounts of computation.
One class of cells responds to dark areas (red), whereas another reacts
to light regions (green). A different set of cells tracks leading edges of
objects (yellow) and trailing edges (blue). The gray-scale images,
generated by decoding these messages, show what a blind person would
see with neuromorphic retinal implants.
or to low-cost, highly effective visual, audio or olfactory recognition chips for robots and other smart machines [see box
on opposite page].
Our team at the University of Pennsylvania initially focused on morphing the retina — the half-millimeter-thick
sheet of tissue that lines the back of the eye. Comprising five
specialized layers of neural cells, the retina “preprocesses”
incoming visual images to extract useful information without
Overview/Inspired by Nature
n
n
n
58
Today’s computers can perform billions of operations
per second, but they are still no match for even a young
child when it comes to skills such as pattern recognition
or visual processing. The human brain is also millions of
times more energy-efficient and far more compact than
a typical personal computer.
Neuromorphic microchips, which take cues from neural
structure, have already demonstrated impressive
power reductions. Their efficiency may make it possible
to develop fully implantable artificial retinas for people
afflicted by certain types of blindness as well as better
electronic sensors.
Someday neuromorphic chips could even replicate the
self-growing connections the brain uses to achieve its
amazing functional capabilities.
SCIENTIFIC A MERIC A N
t h e n e a r ly o n e m i l l i o n ganglion cells in the retina
compare visual signals received from groups of half a dozen to
several hundred photoreceptors, with each group interpreting
what is happening in a small portion of the visual field. As
features such as light intensity change in a given sector, each
ganglion cell transmits pulses of electricity (known as spikes)
along the optic nerve to the brain. Each cell fires in proportion
to the relative change in light intensity over time or space— not
to the absolute input level. So the nerve’s sensitivity wanes with
growing overall light intensity to accommodate, for example,
the five-decade rise in the sky’s light levels observed from predawn to high noon.
Misha A. Mahowald, soon after earning her undergraduate biology degree, and Carver Mead, the renowned microelectronics technologist, pioneered efforts to reproduce the
retina in silicon at the California Institute of Technology. In
their groundbreaking work, Mahowald and Mead reproduced
the fi rst three of the retina’s five layers electronically [“The
Silicon Retina,” by Misha A. Mahowald and Carver Mead;
Scientific American, May 1991]. Other researchers, several of whom passed through Mead’s Caltech laboratory (the
author included), have morphed succeeding stages of the visual system as well as the auditory system. Kareem Zaghloul
morphed all five layers of the retina in 2001 when he was a
doctoral student in my lab, making it possible to emulate the
visual messages that the ganglion cells, the retina’s output neurons, send to the brain. His silicon retina chip, Visio1, replicates responses of the retina’s four major types of ganglion
cells, which feed into and together make up 90 percent of the
optic nerve [see illustration on this page].
Zaghloul represented the electrical activity of each neuron
in the eye’s circuitry by an individual voltage output. The voltage controls the current that is conveyed by transistors connected between a given location in the circuit and other points,
mimicking how the body modulates the responses of neural
synapses. Light detected by electronic photosensors affects the
voltage in that part of the circuit in a way that is analogous to
how it affects a corresponding cell in the retina. And by tiling
copies of this basic circuit on his chip, Zaghloul replicated the
activity in the retina’s five cell layers [see box on page 60].
The chip emulates the manner in which voltage-activated
ion channels cause ganglion cells (and neurons in the rest of the
brain) to discharge spikes. To accomplish this, Zaghloul installed transistors that send current back onto the same location
in the circuit. When this feedback current arrives, it increases
the voltage further, which in turn recruits more feedback current and causes additional amplification. Once a certain initial
level is reached, this regenerative effect accelerates, taking the
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
M AY 2 0 0 5
S T O N E ( e y e p h o t o g r a p h) ; K A R E E M Z A G H L O U L ( r e t i n a l a y o u t) ; J E N C H R I S T I A N S E N ( p h o t o i l l u s t r a t i o n) ( p r e c e d i n g p a g e s) ;
K A R E E M Z A G H L O U L ( t h i s p a g e)
Neuromorphing the Retina
NEUROMORPHIC ELECTRONICS RESEARCH GROUPS
Researchers seek to close the efficiency gap between electronic sensors and the body’s neural networks with microchips that
emulate the brain. This work focuses on small sensor systems that can be implanted in the body or installed in robots.
ORGANIZATION
INVESTIGATORS
PRINCIPAL OBJECTIVES
Johns Hopkins
University
Andreas Andreou,
Gert Cauwenberghs,
Ralph Etienne-Cummings
Battery-powered speech recognizer, rhythm generator for locomotion and
camera that extracts object features
ETH Zurich
(University of Zurich)
Tobi Delbruck, Shi-Chii Liu,
Giacomo Indiveri
Silicon retina and attention chip that automatically select salient regions
in a scene
University of Edinburgh
Alan Murray, Alister Hamilton
Artificial noses and automatic odor recognition based on timing of signaling spikes
Georgia Institute
of Technology
Steve DeWeerth, Paul Hasler
Coupled rhythm generators that coordinate a multisegmented robot
HKUST, Hong Kong
Bertram Shi
Binocular processor for depth perception and visual tracking
Massachusetts Institute
of Technology
Rahul Sarpeshkar
Cochlea-based sound processor for implants for deaf patients
University of Maryland
Timothy Horiuchi
Sonar chip modeled on bat echolocation
University of Arizona
Charles Higgins
Motion-sensing chip based on fly vision
voltage all the way to the highest level, resulting in a spike.
At 60 milliwatts, Zaghloul’s neuromorphic chip uses 1,000
times less electricity than a PC. With its low power needs, this
silicon retina could pave the way for a total intraocular prosthesis — with camera, processor and stimulator all implanted
inside the eye of a blind person who has retinitis pigmentosa
or macular degeneration, diseases that damage photoreceptors
but spare the ganglion cells. Retinal prostheses currently being
developed, for example at the University of Southern California, provide what is called phosphene vision— recipients perceive the world as a grid of light spots, evoked by stimulating
the ganglion cells with microelectrodes implanted inside the
eye — and require a wearable computer to process images captured by a video camera attached to the patient’s glasses. Because the microelectrode array is so small (fewer than 10 pixels by 10 pixels), the patient experiences tunnel vision— head
movements are needed to scan scenes.
Alternatively, using the eye itself as the camera would solve
the rubbernecking problem, and our chip’s 3,600 ganglion-cell
outputs should provide near-normal vision. Biocompatible encapsulation materials and stimulation interfaces, however, need
further refinement before a high-fidelity prosthesis becomes a
reality, maybe by 2010. Better understanding of how various
retinal cell types respond to stimulation and how they contribute to perception is also required. In the interim, such neuromorphic chips could find use as sensors in automotive or security applications or in robotic or factory automation systems.
Metamorphing Neural Connections
t h e p ow e r s av i ng s we attained by morphing the retina
were encouraging, a result that started me thinking about how
the brain actually achieves high efficiency. Mead was prescient
when he recognized two decades ago that even if computing
managed to continue along the path of Moore’s law (which
states that the number of transistors per square inch on intew w w. s c ia m . c o m
grated circuits doubles every 18 months), computers as we
know them could not reach brainlike efficiency. But how could
this be accomplished otherwise? The solution dawned on me
eight years ago.
Efficient operation, I realized, comes from the degree to
which the hardware is customized for the task at hand. Conventional computers do not allow such adjustments; the software is tailored instead. Today’s computers use a few generalpurpose tools for every job; software merely changes the order
in which the tools are used. In contrast, customizing the hardware is something the brain and neuromorphic chips have in
common— they are both programmed at the level of individual connections. They adapt the tool to the specific job. But how
does the brain customize itself? If we could translate that
mechanism into silicon— metamorphing— we could have our
neuromorphic chips modify themselves in the same fashion.
Thus, we would not need to painstakingly reverse-engineer the
brain’s circuits. I started investigating neural development,
hoping to learn more about how the body produces exactly the
tools it needs.
Building the brain’s neural network— a trillion (1012) neurons connected by 10 quadrillion (1016) synapses — is a daunting task. Although human DNA contains the equivalent of a
billion bits of information, that amount is not sufficient to specify where all those neurons should go and how they should connect. After employing its genetic information during early development, the brain customizes itself further through internal
interactions among neurons and through external interactions
with the world outside the body. In other words, sensory neurons wire themselves in response to sensory inputs. The overall
rule that regulates this process is deceptively simple: neurons
that fire together wire together. That is, out of all the signals that
a neuron receives, it accepts those from neurons that are consistently active when it is active, and it ignores the rest.
To learn how one layer of neurons becomes wired to an-
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
SCIENTIFIC A MERIC A N
59
60
KWABENA BOAHEN is a neuromorphic engineer and associate
professor of bioengineering at the University of Pennsylvania.
He left his native Ghana to pursue undergraduate studies in
electrical and computer engineering at Johns Hopkins University in 1985 and became interested in neural networks soon
thereafter. Boahen sees a certain elegance in neural systems
that is missing in today’s computers. He seeks to capture this
sophistication in his silicon designs.
SCIENTIFIC A MERIC A N
RETINAL
Biological sensory systems provide compact, energy-efficient
models for neuromorphic electronic sensors. Engineers
attempting to duplicate the retina in silicon face a
tough challenge: the retina is only half
a millimeter thick, weighs half a gram
CROSS SECTION OF EYE
and consumes the equivalent of just a
tenth of a watt of power. Recent work
Retina
at the University of Pennsylvania
Lens
has yielded a rudimentary
silicon retina.
Optic nerve
CROSS SECTION OF RETINA
Photoreceptors
(rods and cones)
Horizontal cell
Amacrine cell
Bipolar cell
Ganglion cell
Taba designed to conduct charge like a transistor. Charge diffuses through the lattice much like the chemicals released by
tectal cells do through neural tissue. The silicon growth cones
sense this simulated diffusing “chemical” and drag their softwires up the gradient— toward the charge’s silicon neuron
source — by updating the look-up table. Because the charge
must be released by the silicon neuron and sensed by the silicon
growth cone simultaneously, the softwires end up connecting
neurons that are active at the same time. Thus, Neurotrope1
wires together neurons that fire together, as would occur in a
real growing axon.
Starting with scrambled wiring between the Visio1 chip
and the Neurotrope1 chip, Taba successfully emulated the tendency of neighboring retinal ganglion cells to fire together by
activating patches of silicon ganglion cells at random. After
stimulating several thousand patches, he observed a dramatic
change in the softwiring between the chips. Neighboring artificial ganglion cells now connected to neurons in the silicon
tectum that were twice as close as the initial connections. Be-
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
M AY 2 0 0 5
B R YA N C H R I S T I E D E S I G N ; K A R E E M Z A G H L O U L ( s i l i c o n r e t i n a)
THE AUTHOR
other, neuroscientists have studied the frog’s retinotectal projection, which connects its retina to its tectum (the part of the
midbrain that processes inputs from sensory organs). They
have found that wiring one layer of neurons to another occurs
in two stages. A newborn neuron extends projections (“arms”)
in a multilimbed arbor. The longest arm becomes the axon,
the cell’s output wire; the rest serve as dendrites, its input
wires. The axon then continues to grow, towed by an amoeboid structure at its tip. This growth cone, as scientists call it,
senses chemical gradients laid down by trailblazing precursors
of neural communication signals, thus guiding the axon to the
right street in the tectum’s city of cells but not, so to speak, to
the right house.
Narrowing the target down to the right house in the tectum
requires a second step, but scientists do not understand this
process in detail. It is well known, though, that neighboring
retinal ganglion cells tend to fire together. This fact led me to
speculate that an axon could find its retinal cell neighbors in
the tectum by homing in on chemical scents released by active
tectal neurons, because its neighbors were most likely at the
source of this trail. Once the axon makes contact with the tectal neuron’s dendritic arbor, a synapse forms between them and,
voilà, the two neurons that fire together are wired together.
In 2001 Brian Taba, a doctoral student in my lab, built a
chip modeled on this facet of the brain’s developmental process. Because metal wires cannot be rerouted, he decided to
reroute spikes instead. He took advantage of the fact that Zaghloul’s Visio1 chip outputs a unique 13-bit address every time
one of its 3,600 ganglion cells spikes. Transmitting addresses
rather than spikes gets around the limited number of input/
output pins that chips have. The addresses are decoded by the
receiving chip, which re-creates the spike at the correct location in its silicon neuron mosaic. This technique produces a
virtual bundle of axons running between corresponding locations in the two chips — a silicon optic nerve. If we substitute
one address with another, we reroute a virtual axon belonging
to one neuron (the original address) to another location (the
substituted address). We can route these “softwires,” as we
call them, anywhere we want to by storing the substitutions in
a database (a look-up table) and by using the original address
to retrieve them [see box on page 62].
In Taba’s artificial tectum chip, which he named Neurotrope1, softwires activate gradient-sensing circuits (silicon
growth cones) as well as nearby silicon neurons, which are
situated in the cells of a honeycomb lattice. When active, these
silicon neurons release electrical charge into the lattice, which
NEURONS AND NEUROMORPHIC VISION CHIPS
BIOLOGICAL RETINA
SILICON RETINA
The cells in the retina, which are interconnected, extract
information from the visual field by engaging in a complex
web of excitatory (one-way arrows), inhibitory (circles on
a stick), and conductive or bidirectional (two-way arrows)
signaling. This circuitry generates the selective responses
of the four types of ganglion cells (at bottom) that make up
90 percent of the optic nerve’s fibers, which convey visual
information to the brain. On (green) and Off (red) ganglion
cells elevate their firing (spike) rates when the local light
intensity is brighter or darker than the surrounding region.
Inc (blue) and Dec (yellow) ganglion cells spike when the
intensity is increasing or decreasing, respectively.
Neuromorphic circuits emulate the complex
interactions that occur among the various retinal cell
types by replacing each cell’s axons and dendrites
(signal pathways) with metal wires and each synapse
with a transistor. Permutations of this arrangement
produce excitatory and inhibitory interactions that
mimic similar communications among neurons. The
transistors and the wires that connect them are
laid out on silicon chips. Various regions of the chip
surface perform the functions of the different cell
layers. The large green squares are phototransistors,
which transduce light into electricity.
SILICON CHIP DETAIL
Photoreceptor
Conductive
interaction
Horizontal
Bipolar
Excitatory
interaction
Amacrine
Inhibitory
interaction
Ganglion
5 microns
ON
INC
DEC
OFF
cause of noise and variability, however, the wiring was not
perfect: terminals of neighboring cells in the silicon retina did
not end up next to one another in the silicon tectum. We wondered how the elaborate wiring patterns thought to underlie
biological cortical function arise — and whether we could get
further tips from nature to refine our systems.
Cortical Maps
to f i n d ou t, we had to take a closer look at what neuroscience has learned about connections in the cortex, the brain
region responsible for cognition. With an area 16 inches in
diameter, the cortex folds like origami paper to fit inside the
skull. On this amazing canvas, “maps” of the world outside
are drawn during infancy. The best-studied example is what
scientists call area V1 (the primary visual cortex), where visual messages from the optic nerve first enter the cortex. Not
only are the length and width dimensions of an image mapped
onto V1 but also the orientation of the edges of objects therein.
As a result, neurons in V1 respond best to edges oriented at a
w w w. s c ia m . c o m
particular angle — vertical lines, horizontal lines, and so forth.
The same orientation preferences repeat every millimeter or
so, thereby allowing the orientations of edges in different sectors of the visual scene to be detected.
Neurobiologists David H. Hubel and Torsten N. Wiesel,
who shared a Nobel Prize in medicine for discovering the V1
map in the 1960s, proposed a wiring diagram for building a
visual cortex— one that we found intimidating. According to
their model, each cortical cell wires up to two groups of thalamic cells, which act as relays for retinal signals bound for the
cortex. One group of thalamic cells should respond to the sensing of dark areas (which we emulate with Visio1’s Off cells),
whereas the other should react to the sensing of light (like our
Visio1’s On cells). To make a cortical cell prefer vertical edges,
for instance, both groups of cells should be set to lie along a
vertical line but should be displaced slightly so the Off cells lie
just to the left of the On cells. In that way, a vertical edge of an
object in the visual field will activate all the Off cells and all
the On cells when it is in the correct position. A horizontal
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
SCIENTIFIC A MERIC A N
61
edge, on the other hand, will activate only half the cells in each
group. Thus, the cortical cell will receive twice as much input
when a vertical edge is present and respond more vigorously.
At first we were daunted by the detail these wiring patterns
required. We had to connect each cell according to its orientation preference and then modify these wiring patterns systematically so that orientation preferences changed smoothly,
with neighboring cells having similar preferences. As in the
cortex, the same orientations would have to be repeated every
millimeter, with those silicon cells wired to neighboring locations in the retina. Taba’s growth cones certainly could not
cope with this complexity. In late 2002 we searched for a way
to escape this nightmare altogether. Finally, we found an answer in a five-decade-old experiment.
In the 1950s famed computer scientist Alan M. Turing
showed how ordered patterns such as a leopard’s spots or a
cow’s dapples could arise spontaneously from random noise.
We hoped we could use a similar technique to create neighboring regions with similar orientation patterns for our chip. Turing’s idea, which he tested by running simulations on one of
the first electronic computers at the University of Manchester,
was that modeled skin cells would secrete “black dye” or
MAKING CONNECTIONS (BIOLOGICAL OR SILICON)
62
SCIENTIFIC A MERIC A N
ARTIFICIAL RETINA
0
1
2
3
Ganglion cells
Axons
0
1
2
3
0
3
1
0
2
1
3
2
1
2
3
RANDOM-ACCESS MEMORY
0
3
0
1
0
2
2
3
1
Re-created spike position
1
2
3
Tectal cells
Original spike position
0
Electrical charge
Activated cell
ARTIFICIAL TECTUM
Before self-organization
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
BRI A N TA BA A ND JEN CHRIS TI A NSEN
In the early stages of the eye’s
development, ganglion cells in the retina
project axons into a sensory center of the
midbrain called the tectum. The retinal
axons home in on chemical trails released
by neighboring tectal cells that are
activated at the same time, so neurons
that fire together wire together.
Ultimately, a map of the retinal
sensors’ spatial organization forms
in the midbrain.
To emulate this process, University of
Pennsylvania neuromorphic engineers
use “softwires” to self-organize links
between cells in their silicon retina chip,
Visio1 (top), and those in their artificial
tectum chip, Neurotrope1 (bottom).
Electrical output pulses called spikes are
“routed” from the artificial ganglion cells
to the tectal cells using a random-access
memory (RAM) chip (middle). The retinal
chip supplies the address of the spiking
silicon neuron, and the tectal chip recreates that pulse at the corresponding
location. In this example, the artificial
tectum instructs the RAM to swap
address entries 1 and 2. As a result,
ganglion cell 2’s axon terminus moves to
tectal cell 1, bumping ganglion cell 3’s
axon from that location. The axons
“sense” the gradient of electrical
charge released by an activated
silicon tectal cell, which helps to
guide the connections.
After engineers repeatedly activated
patches of neighboring silicon neurons in
the artificial retina (outlined triangles,
top left), the tectal cells’ axon end
points—which were initially widely
distributed (outlined triangles, bottom
left)—grew closer, yielding more uniform
swaths on a colorized map (bottom right).
After self-organization
M AY 2 0 0 5
ORIENTATION PREFERENCES IN THE BRAIN AND IN SILICON
In both the visual cortex of a
ferret (left) and a neuromorphic
cortex chip (right), researchers
have mapped the location of
cells that respond preferentially
to object edges of a certain
orientation (key, below). In
both maps, neighboring cells
tend to have similiar orientation
preferences, which shows that
the cortex chip emulates the
biological system.
M A R C O S F R A N K ( f e r r e t) ; P A U L M E R O L L A ( c h i p)
FERRET CORTEX
“bleach” indiscriminately. By introducing variations among
the cells so that they produced slightly different amounts of
dye and bleach, Turing generated spots, dapples and even zebralike stripes. These slight initial differences were magnified
by blotting and bleaching to create all-or-nothing patterns. We
wondered if this notion would work for cortical maps.
Four years ago computational neuroscientist Misha Tsodyks and his colleagues at the Weizmann Institute of Science
in Rehovot, Israel, demonstrated that, indeed, a similar process could generate cortexlike maps in software simulations.
Paul Merolla, another doctoral student in my lab, took on the
challenge of getting this self-organizing process to work in
silicon. We knew that chemical dopants (impurities) introduced during the microfabrication process fell randomly,
which introduced variations among otherwise identical transistors, so we felt this process could capture the randomness
of gene expression in nature. That is putatively the source of
variation of spot patterns from leopard to leopard and of orientation map patterns from person to person. Although the
cells that create these patterns in nature express identical
genes, they produce different amounts of the corresponding
dye or ion channel proteins.
With this analogy in mind, Merolla designed a single silicon neuron and tiled it to create a mosaic with neuronlike
excitatory and inhibitory connections among neighbors,
which played the role of blotting and bleaching. When we
fi red up the chips in 2003, patterns of activity— akin to a
leopard’s spots — emerged. Different groups of cells became
active when we presented edges with various orientations. By
marking the locations of these different groups in different
colors, we obtained orientation preference maps similar to
those imaged in the V1 areas of ferret kits [see box above].
Building Brains in Silicon
having morphed the retina’s five layers into silicon, our goal
turned to doing the same to all six of the visual cortex’s layers.
We have taken a fi rst step by morphing layer IV, the cortex’s
input layer, to obtain an orientation preference map in an imw w w. s c ia m . c o m
CORTEX CHIP
mature form. At three millimeters, however, the cortex is five
times thicker than the retina, and morphing all six cortical
layers requires integrated circuits with many more transistors
per unit area.
Chip fabricators today can cram a million transistors and
10 meters of wire onto a square millimeter of silicon. By the
end of this decade, chip density will be just a factor of 10 shy
of cortex tissue density; the cortex has 100 million synapses
and three kilometers of axon per cubic millimeter.
Researchers will come close to matching the cortex in
terms of sheer numbers of devices, but how will they handle
a billion transistors on a square centimeter of silicon? Thousands of engineers would be required to design these highdensity nanotechnology chips using standard methods. To
date, a hundredfold rise in design engineers accompanied the
10,000-fold increase in the transistor count in Intel’s processors. In comparison, a mere doubling of the number of genes
in flies to that of humans enabled evolutionary forces to construct brains with 10 million times more neurons. More sophisticated developmental processes made possible the increased complexity by elaborating on a relatively simple
recipe. In the same way, morphing neural development processes instead of simply morphing neural circuitry holds great
promise for handling complexity in the nanoelectronic systems of the future.
MORE TO EXPLORE
Analog VLSI and Neural Systems. Carver Mead. Addison-Wesley, 1989.
Topographic Map Formation by Silicon Growth Cones. Brian Taba and
Kwabena Boahen in Advances in Neural Information Processing
Systems, Vol. 15. Edited by Suzanna Becker, Sebastian Thrun and
Klaus Obermayer. MIT Press, 2003.
Optic Nerve Signals in a Neuromorphic Chip. Kareem A. Zaghloul and
Kwabena Boahen in IEEE Transactions on Biomedical Engineering,
Vol. 51, No. 4, pages 657–675; 2004.
A Recurrent Model of Orientation Maps with Simple and Complex
Cells. Paul Merolla and Kwabena Boahen in Advances in Neural
Information Processing Systems, Vol. 16. Edited by Sebastian Thrun,
Larry Saul and Bernhard Sholkopf. MIT Press, 2004.
The author’s Web site: www.neuroengineering.upenn.edu/boahen
COPYRIGHT 2005 SCIENTIFIC AMERICAN, INC.
SCIENTIFIC A MERIC A N
63
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement