Download A Brain-Like Computer for Cognitive Applications: The

A Brain-Like Computer for
Cognitive Applications:
The Ersatz Brain Project
James A. Anderson
[email protected]
Department of Cognitive and Linguistic Sciences
Brown University, Providence, RI 02912
Paul Allopenna
[email protected]
Aptima, Inc.
12 Gill Street, Suite 1400, Woburn, MA
Our Goal:
We want to build a first-rate, second-rate
brain.
Participants
Faculty:
Jim Anderson, Cognitive Science.
Gerry Guralnik, Physics.
Gabriel Taubin, Engineering.
Students, Past and Present:
Socrates Dimitriadis, Cognitive Science.
Dmitri Petrov, Physics.
Erika Nesse, Cognitive Science.
Brian Merritt, Cognitive Science.
Staff:
Samuel Fulcomer,
Jim O’Dell, Center for Computation and Visualization.
Private Industry:
Paul Allopenna, Aptima, Inc.
John Santini, Anteon, Inc.
Why Build a Brain-Like Computer?
1. Engineering.
Computers are all special purpose devices.
Many of the most important practical computer applications
of the next few decades will be cognitive in nature:





Natural language processing.
Internet search.
Cognitive data mining.
Decent human-computer interfaces.
Text understanding.
We feel it will be necessary to have a cortex-like
architecture (either software or hardware) to run these
applications efficiently.
2. Science:
Such a system, even in simulation, becomes a
powerful research tool.
It leads to designing models with a particular
structure to match the brain-like computer.
If we capture any of the essence of the cortex,
writing good programs will give insight into
the biology and cognitive science.
If we can write good software for a vaguely brain
like computer we may show we really understand
something important about the brain.
3. Personal:
It would be the ultimate cool gadget.
A technological vision:
In 2050 the personal computer you buy in Wal-Mart will
have two CPU’s with very different architecture:
First, a traditional von Neumann machine that runs
spreadsheets, does word processing, keeps your
calendar straight, etc. etc. What they do now.
Second, a brain-like chip

To handle the interface with the von Neumann
machine,

Give you the data that you need from the Web or
your files (but didn’t think to ask for).

Be your silicon friend, guide, and confidant.
History : Technical Issues
Many have proposed the construction of brain-like
computers.
These attempts usually start with

massively parallel arrays of neural computing
elements

elements based on biological neurons, and

the layered 2-D anatomy of mammalian cerebral cortex.
Such attempts have failed commercially.
The early connection machines from Thinking
Machines,Inc.,(W.D. Hillis, The Connection Machine,
1987) was most nearly successful commercially and is
most like the architecture we are proposing here.
Consider the extremes of computational brain models.
First Extreme: Biological Realism
The human brain is composed of on the order of 1010
neurons, connected together with at least 1014 neural
connections. (Probably underestimates.)
Biological neurons and their connections are extremely
complex electrochemical structures. The more
realistic the neuron approximation the smaller the
network that can be modeled.
There is good evidence that for cerebral cortex a
bigger brain is a better brain.
Projects that model neurons are of scientific interest.
They are not large enough to model or simulate
interesting cognition.
Neural Networks.
The most successful brain
inspired models are
neural networks.
They are built from simple
approximations of
biological neurons:
nonlinear integration of
many weighted inputs.
Throw out all the other
biological detail.
Neural Network Systems
Units with these
approximations can build
systems that

can be made large,

can be analyzed,

can be simulated,

can display complex
cognitive behavior.
Neural networks have been
used to model important
aspects of human
cognition.
Second Extreme: Associatively
Linked Networks.
The second class of brain-like
computing models is a basic
part of computer science:
Associatively linked
structures.
One example of such a
structure is a semantic
network.
Such structures underlie most
of the practically
successful applications of
artificial intelligence.
Associatively Linked Networks (2)
The connection between the biological nervous system
and such a structure is unclear.
Few believe that nodes in a semantic network correspond
in any sense to single neurons or groups of neurons.
Physiology (fMRI) suggests that a complex cognitive
structure – a word, for instance – gives rise to
widely distributed cortical activation.
Virtue of Linked Networks:
connected nodes.
They have sparsely
In practical systems, the number of links converging on
a node range from one or two up to a dozen or so.
The Ersatz Brain Approximation:
The Network of Networks.
Received wisdom has it that neurons are the basic
computational units of the brain. The Ersatz Brain
Project is based on a different assumption.
The Network of Networks model was developed in
collaboration with Jeff Sutton (Harvard Medical
School, now NSBRI).
Cerebral cortex contains intermediate level structure,
between neurons and an entire cortical region.
Examples of intermediate structure are cortical
columns of various sizes (mini-, plain, and hyper)
Intermediate level brain structures are hard to study
experimentally because they require recording from
many cells simultaneously.
Cortical Columns: Minicolumns
“The basic unit of cortical operation is the
minicolumn … It contains of the order
of 80-100 neurons except in the
primate striate cortex, where the
number is more than doubled. The
minicolumn measures of the order of
40-50 m in transverse diameter,
separated from adjacent minicolumns
by vertical, cell-sparse zones … The
minicolumn is produced by the
iterative division of a small number of
progenitor cells in the
neuroepithelium.” (Mountcastle, p. 2)
VB Mountcastle (2003). Introduction [to a special
issue of Cerebral Cortex on columns]. Cerebral
Cortex, 13, 2-4.
Figure: Nissl stain of cortex in planum
temporale.
Columns: Functional
Groupings of minicolumns seem to form the
physiologically observed functional columns.
Best known example is orientation columns in
V1.
They are significantly bigger than minicolumns,
typically around 0.3-0.5 mm.
Mountcastle’s summation:
“Cortical columns are formed by the binding together of many
minicolumns by common input and short range horizontal connections.
… The number of minicolumns per column varies … between 50 and
80. Long range intracortical projections link columns with similar
functional properties.” (p. 3)
Cells in a column ~ (80)(100) = 8000
Sparse Connectivity
The brain is sparsely connected. (Unlike most neural
nets.)
A neuron in cortex may have on the order of 100,000
synapses. There are more than 1010 neurons in the
brain. Fractional connectivity is very low: 0.001%.
Implications:
• Connections are expensive biologically since they
take up space, use energy, and are hard to wire up
correctly.
• Therefore, connections are valuable.
• The pattern of connection is under tight control.
• Short local connections are cheaper than long ones.
Our approximation makes extensive use of local
connections for computation.
Network of Networks Approximation
We use the Network of
Networks [NofN]
approximation to structure
the hardware and to reduce
the number of connections.
We assume the basic
computing units are not
neurons, but small (104
neurons) attractor
networks.
Basic Network of Networks
Architecture:
• 2 Dimensional array of
modules
• Locally connected to
neighbors
The activity of the nonlinear attractor
networks (modules) is
dominated by their
attractor states.
Attractor states may be
built in or acquired
through learning.
We approximate the
activity of a module
as a weighted sum of
attractor states.That
is: an adequate set of
basis functions.
Activity of Module:
x = Σ ciai
where the ai are the
attractor states.
Elementary Modules
The Single Module: BSB
The attractor
network we
use for the
individual
modules is
the BSB
network
(Anderson,
1993).
It can be
analyzed
using the
eigenvectors
and
eigenvalues
of its local
connections.
Interactions between Modules
Interactions between modules are described by state
interaction matrices, M. The state interaction matrix
elements give the contribution of an attractor state in
one module to the amplitude of an attractor state in a
module connected to it.
In the linear region
x(t+1) =Σ Msi
+
f
weighted sum
input
from other modules
+
x(t)
ongoing
activity
The Linear-Nonlinear Transition
The first processing stage is linear and sums
influences from other modules.
The second processing stage (with limited values) is
nonlinear.
The linear to nonlinear transition is a powerful
computational tool for cognitive applications.
It also describes the processing path taken by many
cognitive processes.
Generalization from cognitive science:
Sensory inputs  (categories, concepts, words)
Processing moves from continuous values to discrete
entities. (McCulloch and Pitts had it backwards.)
Binding Module Patterns Together.
An associative Hebbian
learning event will tend
to link f with g through
the local connections.
There is a speculative
connection to the
important binding
problem of cognitive
science and
neuroscience.
The larger groupings will
act like a unit.
Responses will be stronger
to the pair f,g than to
either f or g by itself.
Two adjacent modules interacting.
Hebbian learning will tend to bind
responses of modules together if f
and g frequently co-occur.
We can extend this
associative binding model
to larger scale groupings.
It may become possible to
suggest a natural way to
bridge the gap in scale
between single neurons and
entire brain regions.
Networks >
Networks of Networks >
Networks of
(Networks of Networks) >
Networks of
(Networks of (Networks
of Networks))
and so on …
Scaling
Interference Patterns
We are using local transmission of (vector)
patterns, not scalar activity level.
We have the potential for traveling pattern waves
using the local connections.
This lateral information flow allows the potential
for the formation of feature combinations in the
interference patterns where two different
patterns col
Learning the Interference Pattern
The individual modules are nonlinear learning networks.
We can form new attractor states when an interference
pattern forms when two patterns meet at a module.
Module Evolution
Module evolution with learning:

From an initial repertoire of basic attractor
states

to the development of specialized pattern
combination states unique to the history of
each module.
Geometry of Interference Patterns
Pattern information travels laterally. Patterns
converge on particular locations.
Some spatial (topographic) patterns of module
activation should be favored by NofN learning.
X
Examples: X --- X --- X
/ \
X---X
These equal distance arrangements give good
convergence.
The topographic arrangement of the data and the
computation becomes critical.
“Topographic programming” becomes a potential
useable feature of the software.
Biological Evidence:
Columnar Organization in IT
Tanaka (2003)
suggests a columnar
organization of
different response
classes in primate
inferotemporal
cortex.
There seems to be
some internal
structure in these
regions: for
example, spatial
representation of
orientation of the
image in the
column.
IT Response Clusters: Imaging
Tanaka (2003) used
intrinsic visual
imaging of cortex.
Train video camera
on exposed cortex,
cell activity can
be picked up.
At least a factor of
ten higher
resolution than
fMRI.
Size of response is
around the size of
functional columns
seen elsewhere:
300-400 microns.
Columns: Inferotemporal Cortex
Responses of a region
of IT to complex
images involve
discrete columns.
The response to a
picture of a fire
extinguisher shows
how regions of
activity are
determined.
Boundaries are where
the activity falls
by a half.
Note: some spots are
roughly equally
spaced.
Active IT Regions for a Complex Stimulus
Note the large number of roughly equally distant
spots (2 mm) for a familiar complex image.
Histogram of Distances
Were able to plot
histograms of
distances in a number
of published IT
intrinsic images of
complex figures.
Distances computed from
data in previous
figure (Dimitriadis)
Generalization
Simple transformations of some complex images
(here rotation of a face) are stored in
adjacent cortical locations.
Note the smooth translation of activity along
the cortical surface.
Revised Columnar Structure
Tanaka suggested this
might be general.
Implications: Area TE
in IT stores “theme
plus variations”,
that is, an image
plus its most common
and natural
transformations.
Generalization is hard:
Here are “regions”
representing useful
generalizations.
Network of Networks Functional Summary.
• The NofN approximation assumes a two dimensional array of
attractor networks.
• The attractor states dominate the output of the system at
all levels.
• Interactions between different modules are approximated by
interactions between their attractor states.
• Lateral information propagation plus nonlinear learning
allows formation of new attractors at the location of
interference patterns.
• There is a linear and a nonlinear region of operation in
both single and multiple modules.
• The qualitative behavior of the attractor networks can be
controlled by analog gain control parameters.
Engineering Hardware Considerations
We feel that there is a size, connectivity, and computational
power “sweet spot” at the level of the parameters of the
network of network model.
If an elementary attractor network has 104 actual neurons,
that network display 50 attractor states. Each elementary
network might connect to 50 others through state
connection matrices.
A brain-sized system might consist of 106 elementary units
with about 1011 (0.1 terabyte) numbers specifying the
connections.
If 100 to 1000 elementary units can be placed on a chip there
would be a total of 1,000 to 10,000 chips in a cortex
sized system.
These numbers are large but within the upper bounds of
current technology.
Proposed
Basic System
Architecture
Our basic computer architecture consists of

a potentially huge (millions) number

of simple CPUs

connected locally to each other and

arranged in a two dimensional array.

The (sparse) longer range connections are simulated in
software.
We assume each CPU can be identified with a single attractor
network in the Network of Networks model.
A Software Example:
Sensor Fusion
One potential application is to sensor fusion. Sensor fusion
means merging information from different sensors into a
unified interpretation.
Involved in such a project in collaboration with Texas
Instruments and Distributed Data Systems, Inc.
The project was a way to do the de-interleaving problem in
radar signal processing using a neural net.
In a radar environment the problem is to determine how many
radar emitters are present and whom they belong to.
Biologically, this corresponds to the behaviorally important
question, “Who is looking at me?” (To be followed, of
course, by “And what am I going to do about it?”)
Radar
A receiver for radar pulses provide several kinds of
quantitative data:
•
•
•
•
•
frequency,
intensity,
pulse width,
angle of arrival, and
time of arrival.
The user of the radar system wants to know qualitative
information:
•
•
•
•
How many emitters?
What type are they?
Who owns them?
Has a new emitter appeared?
Concepts
The way we solved the problem was by using a
concept forming model from cognitive science.
Concepts are labels for a large class of members
that may differ substantially from each other.
(For example, birds, tables, furniture.)
We built a system where a nonlinear network
developed an attractor structure where each
attractor corresponded to an emitter.
That is, emitters became discrete, valid
concepts.
Human Concepts
One of the most useful computational properties
of human concepts is that they often show a
hierarchical structure.
Examples might be:
animal > bird > canary > Tweetie
or
artifact > motor vehicle > car > Porsche > 911.
A weakness of the radar concept model is that it
did not allow development of these important
structures.
Sensor Fusion with the Ersatz Brain.
We can do simple sensor fusion in the Ersatz Brain.
The data representation we develop is directly based on
the topographic data representations used in the
brain: topographic computation.
Spatializing the data, that is letting it find a
natural topographic organization that reflects the
relationships between data values, is a technique of
great potential power.
Spatializing the problem provides a way of
“programming” a parallel computer.
Topographic Data Representation
Low Values
Medium Values
High Values
••++++••••••••••••••••••••••••••••••••••••••••••••
•••••••••••••••••••••••++++•••••••••••••••••••••••
••••••••••••••••••••••••••••••••••••••••••••++++••
We initially will use a simple bar code to code the value of
a single parameter.
The precision of this coding is low.
This loss of precision disturbed traditional radar engineers:
we deliberately threw out their hard won precision.
But we didn’t care about quantitative precision:
qualitative analysis.
We wanted
For our demo Ersatz
Brain program, we
will assume we
have four
parameters derived
from the source.
An “object” is
characterized by
values of these
four parameters,
coded as bar codes
on the edges of
the array of CPUs.
We assume local
linear
transmission of
patterns from
module to module.
Demo
Each pair of
input patterns
gives rise to
an interference
pattern, a line
perpendicular
to the midpoint
of the line
between the
pair of input
locations.
There are places
where three or four
features meet at a
module.
The higher-level
combinations
represent relations
between the
individual data
values in the input
pattern.
The higher level
combinations have
literally fused
spatial relations
of the input data,
Formation of Hierarchical Concepts.
This approach allows the formation of what look like
hierarchical concept representations.
Suppose we have three parameter values that are fixed for
each object and one value that varies widely from
example to example.
The system develops two different types of spatial data.
In the first, some high order feature combinations are
fixed since the three fixed input (core) patterns never
change.
In the second there is a varying set of feature
combinations corresponding to the details of each
specific example of the object.
The specific examples all contain the common core pattern.
Core Representation
The group of
coincidences
in the
center of
the array is
due to the
three input
values
arranged
around the
left, top
and bottom
edges.
Left are two examples where
there is a different
value on the right side
of the array. Note the
common core pattern
(above).
Development of A “Hierarchy” Through
Spatial Localization.
The coincidences due to the core (three values) and to
the examples (all four values) are spatially
separated.
We can use the core as a representation of the examples
since it is present in all of them.
The core represents relations between the data values,
not the data itself.
It acts as the higher level in a simple hierarchy: all
examples contain the core.
The many-to-one relationship here – many low level
examples, fewer high level examples -- is typical of
a hierarchical semantic networks.
Conclusions
The Ersatz Brain Project has led us down an
interesting path.
If we start to require software to use brain-like
constraints,then new ways to tackle old
problems emerge.
• New “analog” control structures: We can use
spatial “programming patterns” to do
arithmetic.
• We can spatialize the computation, the data and
the solutions through initial representations
and feature combinations.
Conclusions
• Potential emergence of hierarchical structure.
• We can use related techniques to do
disambiguation using context and semantic
networks.
These ideas might be of value for current
computers.
I feel that their real domain of application will
be to the computers of the future.