The human brain is far too complex, the biggest brain simulator built by IBM, run by a 147,456-processor supercomputer, can only reach the level of a cat’s brain.
This brain simulator was developed by IBM’s Almaden research center. It consists of 1.6 billion virtual neurons connected by 9 trillion synapses, which is almost a cell-by-cell simulation of the human visual cortex.
With all the massive components, over 150,000 gigabytes of memory and power consumption of over million watts, the simulator is only as good as a cat’s brain. But the achievement is much better than the smaller rat brain simulator that they made two years ago.
To simulate the entire human cortex, they need computing power that is 1,000 times faster than what they currently have managed, but this could only be achieved in about a decade from now.
via gizmowatch
November 22nd, 2009 at 12:21 am
[...] 20. IBM’s biggest brain simulator works as good as a cat’s brain (only) [...]
November 25th, 2009 at 5:08 pm
BM’s claim is a HOAX.
This is a mega public relations stunt – a clear case of scientific deception
of the public. These simulations do not even come close to the complexity of
an ant, let alone that of a cat. IBM allows Mohda to mislead the public into
believing that they have simulated a brain with the complexity of a cat -
sheer nonsense.
Here are the scientific reasons why this is a hoax and misleading PR stunt:
How complex is their model?
They claim to have simulated over a billion neurons interacting. Their so
called “neurons” are the tiniest of points you can imagine, a microscopic
dot. Over 98% of the volume of a neuron is branches (like a tree). They just
cut off all the branches and roots and took a point in the middle of the
trunk to represent a entire neuron. In real life, each segment of the
branches of a neuron contains dozens of ion channels that powerfully
controls the information processing in a neuron. They have none of that.
Neurons contain 10′s of thousands of proteins that form a network with 10′s
of millions of interactions. These interactions are incredibly complex and
will require solving millions of differential equations. They have none of
that. Neurons contain around 20’000 genes that produce products called mRNA,
which builds the proteins. The way neurons build proteins and transport them
to all the corners of the neuron where they are needed is an even more
complex process which also controls what a neuron is, its memories and how
it will process information. They have none of that. They use an alpha
function (up fast down slow) to simulate a synaptic event. This is a
completely inaccurate representation of a synapse. There are at least 6
types of synapses that are highly non-linear in their transmission (i.e.
that transform inputs and not only transmit inputs). In fact you would need
a 10′s of thousands of differential equations to simulate one synapse.
Synapses are also extremely complex molecular machines that would themselves
require thousands of differential equations to simulate just one. They
simulated none of this. There are complex differential equations that must
be solved to simulate the ionic flow in the branches, to simulate the ion
channels biophysics, the protein-protein interactions, as well as the
complete biochemical and genetic machinery as well as the synaptic
transmission between neurons. 100′s of thousands of more differential
equations. They have none of this. Then there are glia – 10 times more than
neurons..And the blood supply…and more and more. These “points” they
simulated and the synapses that they use for communication are literally
millions of times simpler than a real cat brain. So they have not even
simulated a cat’s brain at more than one millionth of it’s complexity.
Is it nonetheless the biggest point neuron simulation ever run?
No. These people simulated 1 billion points interacting. They used a
formulation to model the summing up and threshold spiking of the “points”
called the Izhikevik Formulation (an extremely simple equation). Eugene
Izhikevik himself already in 2005 ran a simulation with 100 billion such
points interacting just for the fun of it: (over 60 times larger than
Modha’s simulation). This simulation ran on a cluster of desktop PCs and
which every graduate student can run This is no technical achievement and
certainly not even a record number of point neurons. That model exhibited
oscillations, but that always happens so even simulating 100 Billion such
points interacting is light years away from a brain.
see: http://www.izhikevich.org/human_brain_simulation/Blue_Brain.htm#Simulation%20of%20Large-Scale%20Brain%20Models
Is the simulator they built a big step?
Not even close. There are numerous proprietary and peer-reviewed
neurosimulators (e.g., NCS, pNEURON, SPLIT, NEST) out there that can handle
very large parallel models that are essentially only bound by the available
memory. The bigger the machine you have available, the more neurons you can
simulate. All these simulators apply optimizations for the particular
platform in order to make optimal use of the available hardware. Without any
comparison to existing simulators, their publication is a non-peer reviewed
claim.
Did they learn anything about the brain?
They got very excited because they saw oscillations. Oscillations are an
obligatory artifact that one always gets when many points interact. These
findings that they claim on the neuroscience side may excite engineers, but
not neuroscientists.
Why did they get the Gordon Bell Prize?
They submitted a non-peer reviewed paper to the Gordon Bell Committee and
were awarded the prize almost instantly after they made their press release.
They seem to have been very successful in influencing the committee with
their claim, which technically is not peer-reviewed by the respective
community and is neuroscientifically outrageous.
But is there any innovation here?
The only innovation here is that IBM has built a large supercomputer – which
is irrelevant to the press release.
Why did IBM let Mohda make such a deceptive claim to the public?
I don’t know. Perhaps this is a publicity stunt to promote their
supercompter. The supercomputer industry is suffering from the financial
crisis and they probably are desperate to boost their sales. It is so
disappointing to see this truly great company allow the deception of the
public on such a grand scale.
But have you not said you can simulate the Human brain in 10 years?
I am a biologist and neuroscientist that has studied the brain for 30 years.
I know how complex it is. I believe that with the right resources and the
right strategy it is possible. We have so far only simulated a small part of
the brain at the cellular level of a rodent and I have always been clear
about that.
Would other neuroscientists agree with you?
There is no neuroscientist on earth that would agree that they came even
close to simulating the cat’s brain – or any brain.
But did Mohda not collaborate with neuroscientists?
I would be very surprised if any neuroscientists that he may have had in his
DARPA consortium realized he was going to make such an outrages claim. I
can’t imagine that that the San Fransisco neuroscientists knew he was going
to make such a stupid claim. Modha himself is a software engineer with no
knowledge of the brain.
But did you not collaborate with IBM?
I was collaborating with IBM on the Blue Brain Project at the very beginning
because they had the best available technology to faithfully allow us to
integrate the diversity and complexity found in brain tissue into a model.
This for me is a major endeavor to advance our insights into the brain and
drug development. Two years ago, when the same Dharmendra Mhoda claimed the
“mouse-scale simulations”, I cut all neuroscience collaboration with IBM
because this is an unethical claim and it deceives the public.
What IBM allowed Modha to do here is not only wrong, but outrageous. They
deceived millions of people.
Henry Markram
Blue Brain Project
April 2nd, 2010 at 9:22 am
LOGICAL EXTRACTION OF NEOCORTEX STRUCTURE
I do not understand why the neocortex is a mystery to everyone. Its neuron net circuit is repeated throughout the cortex. It consists of excitatory and inhibitory neurons whose functions, each, have been known for decades. The neuron net circuit is repeated over layers whose axonal outputs feed on as inputs to other layers. The neurons of each layer, each receive axonal inputs from one or more sending layers and all that they can do is correlate the axonal input stimulus pattern with their axonal connection pattern from those inputs and produce an output frequency related to the resultant psps. Axonal growth toward a neuron is definitely the mechanism for permanent memory formation and it is just what is needed to implement conditioned reflex learning. This axonal growth must be under the control of the glial cells and must be a function of the signals surrounding the neurons. However, neurons need to do normalized cross correlations and need to recognize patterns. This requires the inclusion of the inhibitory neuron in the structure to complete the definition of the neocortex.
The cortex is known to be able to do pattern recognition and the correlation between an axonal input stimulus and an axonal input connection pattern is just what is needed to do pattern recognition. However, pattern recognition needs normalized correlations and a means to compare these correlations so that the largest correlation is recognized by the neurons. Without normalization, the psps’ relative values would not be bounded properly and could not be used to determine the best pattern match. In order to get psps to be compared so that the maximum psp neuron would fire, the inhibitory neuron is needed. By having a group of excitatory neurons feed an inhibitory neuron that feeds back inhibitory axonal signals to those excitatory neurons, one is able to have the psps of the excitatory neurons compared, with the neuron with the largest psps firing before the other do as the inhibitory signal decays after each excitatory stimulus, thus inhibiting the other excitatory neurons with the smaller psps. This inhibitory neuron is needed in order to achieve psp comparisons, no question about it. For a meaningful comparison, the psps must be normalized. As unlikely as it may seem possible, it comes out that the inhibitory connections growing by the same rules as excitatory connections, grow to a value which accomplishes the normalization. That is, as the excitatory axon pattern grows via conditioned reflex rules, the inhibitory axon to each excitatory neuron grows to a value equal to the square root of the sum of the squares of the excitatory connections. This can be shown by a mathematical analysis of a group of mutually inhibiting neurons under conditioned reflex learning. This normalization does not require the neurons to behave different from as known for decades, but rather requires that they interact with an inhibitory neuron as described.
Thus, by simply having the inhibitory neurons receive from neighboring excitatory neuron with large connection strengths where if the excitatory neuron fires, the inhibitory neuron fires and by allowing the inhibitory axonal signals be included with the excitatory axonal input signals to the inputs to those excitatory neurons, the neo-cortex is able to do normalized conditioned reflex pattern recognition as its basic function.
If one thinks about it, layers of mutually inhibiting groups of neurons are all that are needed to explain the neo-cortex functions. The layers of neurons are able to exhibit conditioned reflex behavior between sub-patterns, generating new learned behaviors as observed by the human. With layer to layer feedback, multi-stable behavior of layers of neurons results, forming short term memory patterns that become part of the stimulus to other neurons. With normalized correlations, there is always an axonal input stimulus pattern that will excite every excitatory neuron.
The only way to prove this cortex model is to build a simulator, modeling large nets of neurons and observing human behaviors. Most certainly we will never be able to measure the neuron nets of the cortex due to their small sizes. This means, that projects must be formed that do these simulations and do not waste R&D efforts to try to measure properties of the cortex as the main means to understand the cortex. Certainly the area to area connection scheme is needed, but it likely can be varied, still with intelligence being exhibited. Trials will be needed to determine the initial connection strengths when initiating the simulator. These connections will need to be simple such as non-zero between corresponding neurons of the mutually inhibiting groups.
Axon growth toward pulsing neurons is the likely mechanism for memory alteration. Having neuron axons back away from neurons has no physical basis and it is well known that the number of axons increases with age in the human. Certainly axon connection strengths never become proportional to axon pulsing frequencies, otherwise the nets of neurons will never exhibit permanent past memories, but rather be a function of recent events, only. Glial cells are likely participants to axonal growth control. It is likely that they will inhibit axonal growth physically, unless a chemical falls below a concentration. In particular, this would be when the excitatory stimulus (chemically emitted to a neuron by axons to that neuron) to a cell, falls below a critical level, where the correlation between stimulus and connection pattern falls below a limit. The result of such a rule is that learning would only occur if stimulus patterns are new and don’t match the connection patterns on neurons. The psychological effect would be a curiosity behavior, observed in humans. Also, it would result in old age reduction of ability to learn, also observed in humans.
Progress in understanding how the brain works has been basically non-existent over the last 40 years due to limits in measurement. Progress requires simulation to work out the missing details. I predict that simulation will dominate the future efforts of researchers.
Also, I predict that special purpose hardware will dominate the approach. Using conventional computers to simulate nets of neurons in real-time will go out of style very soon.
Simulation permits an evolution process to arrive upon a successful brain understanding. If a logical conclusion is wrong, simulation will eliminate it. If it is right, simulation will verify it.
Dr. Ronald J. Swallow
rswallow@ptd.net
610 704 0914
March 29th, 2011 at 7:09 am
[...] supercomputers cost double figures or more in millions of USD to build. Such as the fastest supercomputer that was [...]