IBM Research – Almaden presented at Supercomputing 2012 last week the next milestone toward fulfilling the ultimate vision of the DARPA’s cognitive computing program, called Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE), according to Dr. Dharmendra S. Modha, Manager, Cognitive Computing, IBM Research – Almaden.
Announced in 2008, DARPA’s SyNAPSE program calls for developing electronic neuromorphic (brain-simulation) machine technology that scales to biological levels, using a cognitive computing architecture with 1010 neurons (10 billion) and 1014 synapses (100 trillion, based on estimates of the number of synapses in the human brain) to develop electronic neuromorphic machine technology that scales to biological levels.”
A network of neurosynaptic cores derived from
long-distance wiring in the monkey brain: Neuro-synaptic cores are locally
clustered into brain-inspired regions, and each core is represented as an
individual point along the ring. Arcs are drawn from a source core to a
destination core with an edge color defined by the color assigned to the source
core. (Credit: IBM)
Simulating 10 billion neurons and 100 trillion synapses on most powerful supercomputer
IBM says it has now accomplished this milestone with its new “TrueNorth” system running on the world’s second-fastest operating supercomputer, the Lawrence Livermore National Lab (LBNL) Blue Gene/Q Sequoia, using 96 racks (1,572,864 processor cores, 1.5 PB memory, 98,304 MPI processes, and 6,291,456 threads).
IBM and LBNL achieved an unprecedented scale of 2.084 billion neurosynaptic cores* containing 53×1010 (530 billion) neurons and 1.37×1014 (100 trillion) synapses running only 1542 times slower than real time.
“We have not built a biologically realistic simulation of the complete human brain,” explains an abstract of the Supercomputing 2012 (SC12) paper (open-access PDF), selected from the 100 SC12 papers as one of the six finalists for the Best Paper Award. “Computation (‘neurons’), memory (‘synapses’), and communication (‘axons,’ ‘dendrites’) are mathematically abstracted away from biological detail toward engineering goals of maximizing function (utility, applications) and minimizing cost (power, area, delay) and design complexity of hardware implementation.”
Two billion neurosynaptic cores
“Previously, we have demonstrated a neurosynaptic core* and some of its applications,” continues the abstract. “We have also compiled the largest long-distance wiring diagram of the monkey brain. Now, imagine a network with over 2 billion of these neurosynaptic cores that are divided into 77 brain-inspired regions with probabilistic intra-region (“gray matter”) connectivity and monkey-brain-inspired inter-region (“white matter”) connectivity.
“This fulfills a core vision of the DARPA SyNAPSE project to bring together nanotechnology, neuroscience, and supercomputing to lay the foundation of a novel cognitive computing architecture that complements today’s von Neumann machines.”
To support TrueNorth, IBM has developed Compass, a multi-threaded, massively parallel functional simulator and a parallel compiler that maps a network of long-distance pathways in the macaque monkey brain to TrueNorth.
* The IBM-Cornell neurosynaptic core is a key building block of a modular neuromorphic architecture, according to Modha. The core incorporates central elements from neuroscience, including 256 leaky integrate-and-fire neurons, 1024 axons, and 256×1024 synapses using an SRAM crossbar memory. It fits in a 4.2mm square area, using a 45nm SOI process.
http://p9.hostingprod.com/@modha.org/blog/2012/06/building_block_of_a_programmab.html
Questions:
1-
How
many years will take to create a robot to simulated the human brain, like Human
android ?
2-
Do
you think it’s possible to create a human simulation ?
3-
Could
you describe the possible consequences of human brain simulation ?
4-
Do
you think a technology like that can improve human civilization ?
1. In my opinion we are still far away from creating a robot that would be even remotely close to mimicking the actual thinking process of humans. So I'm guessing that I will not see that kind of artificial intelligence in my life... but I think that my grandsons will have this opportunity.
ReplyDelete2. Yes. I think that technically it is possible, but I haven't heard about anyone getting close to discovering an algorithm or creating a model that can map the processes occurring in the brain to an electrical device. So again, I think that we will need some time to get to that point.
3. If anyone would achieve such a thing, this would be a sign of a new era. We could create great "farms" of artificial brains, update their knowledge with the most crucial information concerning specific fields of knowledge (like chemistry, biology, etc.) and then let them search for solutions to the biggest problems of mankind... On the other hand I saw the "Matrix" so personally I'm not a big fan of thinking machines... ;)
4. As I stated in the previous answer, I definitely see almost limitless benefits of having that kind of technology, but I also see a lot of drawbacks that would arise from using it. Let's take IT security for instance - if we had a few artificial brains at our disposal we could crack all of the existing security algorithms (at least the ones that exist today).
According to a lot of articles or newsletters available on the web, like Nature, Kurzweil or other publications such progress is occuring.
DeleteDarpa or IBM would not spend any money on a project that would not make sense.
At the moment, these technologies are developing so fast that probably most likely in the next 20 years we’ll see development of computers succesfuly emulating some of the human brain capacities. The general AI, meaning creating a self-aware machine, may take a little bit longer, for eg. another 20 years. But considering the concequences of Moore Law it appears that such technology should be available by the late 40’s.
I would like to refer to your third paragraph because it contains concrete numbers which are too optimistic (in my opinion of course).
DeleteYou wrote “in the next 20 years we’ll see development of computers successfully emulating some of the human brain capacities”. What do you mean when you wrote that we will emulate SOME things in the next 20 years?
I think that some things (like playing chess) computers do better than people now. But in my opinion there is a huge gap between emulating some human brain activities and emulating human brain. Machine shown in your article is “only” something what we already know but in greater scale.
So I think that in twenty years without spectacular discoveries on other (unknown for me) areas we will achieve possibilities for replacing human brain on next areas but we still will be far from creating human simulation. And according to Moore’s law – maybe we will build machines with comparable power to a human brain, but we still be far from creating human simulation because there still will be areas not described in your article. And according to article about learning in XXI century – maybe then our brain will be quite different than now.
Dear Gregory
DeleteTo answer you consternation and lack of current information on the subject, I would gladly refer you to the newest book by Ray Kurzweil "How to create a mind: the secret of human thought revealed ". this is internationaly recognized work and it should fully answer your doubts.
This comment has been removed by the author.
ReplyDeleteThe topic is indeed fascinating, however I don't believe we are very close to creating an artificial human brain. In my opinion the main problem is not the lack of technology capable of running the appropriate processes, but limited knowledge of all the brain complexities. Probably the main bottleneck in such research is that we cannot easily perform experiments to verify all the theories due to ethical issues (at least I hope we can't). For example we still don't know the detailed reasons for many illnesses. On the other hand, a year ago I have heard about the Blue Brain project at EPFL. Its main goal is building a realistic brain model. This research is based on the rat brain analysis though.
ReplyDeleteTo sum up - to me brain simulation is a very exciting thing, but also a little bit scary. What I am affraid of, is not a catastrophic vision presented in movies like "Terminator" or "Matrix" - but the knowledge behind the model and its possible usages on non-artifical brains.
1- How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteWe are probably not far away from simulating the necessary number of neurons and connections, it is a matter of computing power. That doesn't mean we will then have a conscious, thinking artificial mind. A robot actually making use of a brain to perform tasks is likely to take much longer.
2- Do you think it’s possible to create a human simulation ?
In order to simulate all physiological processes of the human body we would first have to understand them all the way down to the cellular level. With so many things going on simultaneously it would be very difficult.
3- Could you describe the possible consequences of human brain simulation ?
In terms of artificial intelligence the real benefits of a brain simulation come when you can simulate it faster than real time or many copies at the same time. Another option is to improve the brain by making it bigger for example. The consequences would be far reaching, from medicine, to decision making.
4- Do you think a technology like that can improve human civilization ?
It would advance our civilization. Our whole civilization is based on gathering knowledge and passing it on from generation to generation. We are naturally able to do that because of our ability to speak and understand language. We've invented writing and today we have networks of computers which takes this to another level. When we can't keep pace with the flow of new information, relying on technology to process it for us seems inevitable.
1 How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteWhether we are referring to fully simulate human intelligence or only evolving machine based on algorythms? Simulating human intelligence will be really hard if even possible (look down), create a self-learning rule-based machine is possible and is happening more and more quickly - as you can see in the video.
2 Do you think it’s possible to create a human simulation?
I'm not sure - as long as the Turing test passage is possible, robot does not a full understand of a metaphor. Abstract thinking is almost impossible to pass because it requires non-deterministic thinking. If even we simulate a non-deterministic way of thinking, it would be still simulated in a deterministic way.
3 Could you describe the possible consequences of human brain simulation ?
This definitely could be a breakthrough in the natural sciences - mathematics, physics - all of which solving difficult and boring tasks (sorry!).
4 Do you think a technology like that can improve human civilization ?
Yes, although I most wait for a quantum computer.
1. How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteIn my opinion, very soon. :) I think that the problem is not in the technology, but in the
high costs associated with the work and research on such solutions.
There are many "conceptual models" showing the solution of that problem,
but as I wrote above in many cases the implementation costs are huge :)
2. Do you think it’s possible to create a human simulation ?
I think that it is possible to create a human simulation.
We shoul remember that simulation is not purely mathematical or only calculated.
I think that in that simulation it must be a coincidence and a "random" element
such as in the evolution - the fortuity is very important.
3. Could you describe the possible consequences of human brain simulation ?
Yes, maybe We can build a new weapon...
We must remember that just as something will not be useful for army,
the government will give it to the people..
4. Do you think a technology like that can improve human civilization ?
This is the great hope for people, but people should use it to learn not for war...
and the story say it does not work. Such as with the nuclear energy.
The first was a bomb, and then there was the nuclear power plants.
Do you think that we (people) can discover (practically, not only theoretically) something new in such scale without thinking about using it for war?
DeleteThere is mentioned DARPA’s program in presented article, so I can’t imagine that sponsor gave money on that program without thinking about war.
1- How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteI can’t imagine such number of years, because human brain is so complicated and acts on many areas (solving problems, creative thinking, feelings and so on).
2- Do you think it’s possible to create a human simulation ?
No, I don’t. I think we don’t need to simulate a human, we need to simulate human brain only in these areas where we expect support from machines in solving our problems.
3- Could you describe the possible consequences of human brain simulation ?
If we discover an artificial brain we will able to use it instead of our current brain. If we don’t control artificial brain completely it will damage us.
4- Do you think a technology like that can improve human civilization ?
Yes, now we should discover how to use it (apart from playing pong).
This comment has been removed by the author.
ReplyDeleteThank all of you for insighful comments. I recommend you read the new Ray Kurzweil book "How to create a mind: the secret of human thought revealed"
ReplyDelete1- How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteI think that the very first thing in order to create brain simulation is to understand how it works, and we are still far from this.
2- Do you think it’s possible to create a human simulation ?
Everything is possible... also you can always create something but will it have sufficient quality ? I think on the current level of technology no, but I'm not specialist in AI.
3- Could you describe the possible consequences of human brain simulation ?
I could only think of the scenarios from the sci-fi movies when machines could live together with humans.
4- Do you think a technology like that can improve human civilization ?
As every tool can be used to do as much harm as good, it always depends on who and how will use it. But still it gives real potential.
The difference is the hardware as you already said. Which is not such a small thing as you make it out to be. To emulate neural networks of any impressive size takes a lot of computer power and energy.
ReplyDeleteSpecialized chips could potentially do that a lot more efficient. Making room for much faster progress in that field.
Simulating the brain's neuron-and-synapse model in new computing architectures might open new avenues for high-performance, energy-efficient computing systems.
Thus I believe that the scale in model should be architected to compute class primitives implemented in silicon that would involve integrated pattern recognition along with other “tools, techniques and procedures” (TTPs). I believe neurons create / retrieve class structures from your brain memory
Cognitive systems will “sense, learn and better predict the consequences of actions”. Doesn’t the prospect of merging artificial intelligence, while less fashionable than it once was, with cognitive systems open the frontiers even further? Imagine the capabilities of a device that can sense, learn, predict and respond.
1.How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteProbably never. I don’t think that there will be robot which exactly simulate human. At some point there will be religion / lobby / fanatics who will want to destroy /prevent it for various reasons. However, work will be continued on developing androids for performing specific tasks.
2. Do you think it’s possible to create a human simulation ?
I think in 100 years or more there will be technology which could be used to create such possibilities. But no one will create such an android unless someone wishes to pay for it.
3. Could you describe the possible consequences of human brain simulation?
There is a lot of science fiction literature on this subject. I do not think it would be good for society.
4. Do you think a technology like that can improve human civilization ?
Such technology can help in medicine. In addition, other technologies will be developed that can be used to replace human in extremely difficult and dangerous work.
How many years will take to create a robot to simulated the human brain, like Human android?
ReplyDeleteI do not know. I think that those future androids would be very sophisticated but still somehow linear in their behavior.
Do you think it’s possible to create a human simulation?
I believe that not only blood and flash constitutes a human being. You need soul to have a working specimen :). That is why a human simulation would smell artificial from long distance.
Could you describe the possible consequences of human brain simulation?
If you think about a brain simulation as a full AI solution then the consequences could be wide range. On one end we could enrich human civilization with a powerful AI “brothers” but on the other end we could gain lethal enemies.
Do you think a technology like that can improve human civilization?
I understood that IBM’s idea is to create a new type of computer which is mostly concentrated on parallel processing. They want to do this to outdistance their competitors. This can change the speed of data processing and brake some barriers we face now. It might bring some good and some opposite.
1- How many years will take to create a robot to simulated the human brain, like Human android ?
ReplyDeleteIn my opinion first we need to know better our brain and for now we are far away form that.
2- Do you think it’s possible to create a human simulation ?
I think that for know it isn't possible also in the future it can be more complicated, because we use to learn and to do decisions not only our experience and some schemes but also we use some kind abstractions in the process of solving problems and so on. And every simulation is only some kind of programs it is set of procedures. In simulation decision are make base of 0,1 where 0 is false and 1 is true the don't use abstracts.
3- Could you describe the possible consequences of human brain simulation ?
It cane be to situation one where we can live in pace and another is the word of matrix and war:)
4- Do you think a technology like that can improve human civilization ?
It depends how it will be use like almost everything what is new can be use in good way or bad way.For example atom can be use in good way to make cheaper and better energy what improve our civilization but also it can be use do make the bomb so in that way it can also destroy our civilization
1) We will never construct a copy of a human brain, because its much more than a collection of neurons, the same as home is something totally different than a house made of bricks.
ReplyDelete2) The question how to define human simultation? The way it acts, how it moves, physiology? Some of those aspects may be simulated now, but what about intuition? How come we can’t replicate that?
3)No, I see only dangers of that.
4) Not necessarily. There is plenty of science fiction novels playing around with the idea, and I think that the human consciousness, the metaphysical “soul” is necessary for any brain to be human-like.
1. How many years will take to create a robot to simulated the human brain, like Human android?
ReplyDeleteI think it will take ages.
2. Do you think it’s possible to create a human simulation ?
We don't need to create human simulation we just need to have a good excuse for getting money for R&D.
3. Could you describe the possible consequences of human brain simulation ?
Human brain will displace humans out of more and more types of job. That will probably have negative influence on society. Many people will become unemployed. Of course there are also positives like progress in science.
4. Do you think a technology like that can improve human civilization?
It depends on how it will be used.