What it takes
As we look for computers and robots to approach the capabilities of the human mind, we find
that the density of storage in computers is already greater than the density of the human
brain except for three factors. Computer chips consume more power than the human brain.
Computer chips are two-dimensional and the human brain is three-dimensional. The
human brain uses some elements like a memristor and current computers do not use
memristors as yet. As a result, progress is punctuated by three things, lower power, three
dimensionality and the use of memristors. Genetic Programming must also be developed
further so that the robot may learn rather than being programmed.
Electronic two-terminal bistable graphitic memories
Transistors are the basis for electronic switching and memory devices as they exhibit
extreme reliabilities with on/off ratios of 10,000-100,000, and billions of these three-
terminal devices can be fabricated on single planar substrates. On the other hand, two-
terminal devices coupled with a nonlinear current-voltage response can be considered as
alternatives provided they have large and reliable on/off ratios and that they can be
fabricated on a large scale using conventional or easily accessible methods. Here, we report
that two-terminal devices consisting of discontinuous 5-10 nm thin films of graphitic sheets
grown by chemical vapor deposition on either nanowires or atop planar silicon oxide exhibit
enormous and sharp room-temperature bistable current-voltage behavior possessing stable,
rewritable, non-volatile and non-destructive read memories with on/off ratios of up to
10,000,000 and switching times of up to 1 microsecond (tested limit). A
nanoelectromechanical mechanism is proposed for the unusually pronounced switching
behavior in the devices.
- it would increase the amount of storage in a two-dimensional array by a factor of five.
- the new switches can be controlled by two terminals instead of three, as in current chips.
This will allow for practical three dimensional memory layers.
- being essentially a mechanical device, such chips will consume virtually no power when
storing memory
- On/off power ratio of one million to one instead of ten to one for phase change memory
[higher is better]
- James Tour said the new switches are also fast; in fact, they react faster than his lab's
current testing systems can measure.
- they're robust. "We've tested it in the lab 20,000 times with no degradation," said Tour. "Its
lifetime is going to be huge, much better than flash memory."
- Typically, graphene is very hard to think about fabricating commercially," he said, "but this
can be done very easily by deposition.
Taken from Nature: Materials.
3-D memristor chip debutsR. Colin JohnsonEE Times (11/26/2008 10:36 AM EST)
PORTLAND, Ore. — Memristors technology got a boost recently from Hewlett-Packard Labs,
which described the first 3-D memristor chip at a conference in Berkeley, Calif. The
Memristor and Memristive Systems Symposium was co-sponsored by the University of
California, the Semiconductor Industry Association and the National Science Foundation.
HP Labs (Palo Alto, Calif.) provided details of a prototype chip designed by HP researcher
Qiangfei Xia that stacked memristor crossbar memory cells on top of a CMOS logic chip. "Xia
used imprint lithography to add a memristor crossbar on top of a CMOS logic circuit," said HP
Labs Fellow Stan Williams, inventor of HP's memristive memory technology. "He has built an
integrated hybrid circuit with both transistors and memristors." Williams and HP colleague
Greg Snider previously proposed an FPGA in which configuration bits were located above
CMOS transistors in a memristor crossbar. Memristor crossbars include two titanium dioxide
layers between two perpendicular arrays of metal lines. One layer of titanium oxide is doped
with oxygen vacancies, making it a semiconductor. The adjacent layer is undoped, leaving it
in its natural state as an insulator. When a crossbar junction is addressed by
simultaneously applying a voltage to one crossbar line on the top and bottom layers, oxygen
vacancies drift from the doped to the undoped layer. This causes it to begin conducting,
turning "on" the memory bit. The bit can again be turned "off" by changing the current
direction, whereupon oxygen vacancies migrate back into the doped layer. According to
Williams, HP Labs' memristor-based FPGA demonstrates that a CMOS fab can make
integrated memristor/transistor circuits in three dimensions. Also at the symposium, Snider
unveiled a design that used memristors in their analog mode as synapses in a neural
computing architecture. Memristor crossbars are the only technology that is dense enough
to simulate the human brain, Snider claimed, adding that the HP Labs crossbars are ten
times denser than synapses in the human cortex. By stacking crossbars on a CMOS logic
chip, variable resistance could mimic the learning functions of synapses in neural
networks. HP Labs and Boston University were recently awarded a contract by the Defense
Advanced Research Projects Agency to build the first artificial neural network based on
memristors. Also at the conference, Massimiliano Di Ventra of the University of California at
San Diego described how memristors can explain biological learning in amoebas. Amoebas
learn to change their behavior in a manner that can be explained by an LC circuit and a
memristor. Di Ventra also presented evidence that microscopic memristive elements are
present in unicellular as well as multicellular organisms.
iRobis Announces Complete Cognitive Software System for Robots
IT, New Media & Software
Press release from: iRobis
(openPR) - November 27, 2008 - Institute of Robotics in Scandinavia (iRobis) has announced
that the world’s first “complete cognitive software system for robotics” is ready for application.
The system turns robots into self-developing, adaptive, problem-solving, “thinking”
machines.
VP for Business Development Roger F. Gay made the announcement November 19th at the
RoboDevelopment Conference & Expo in Santa Clara, California. Development of the software
system, called Brainstorm™, was previously discussed by lead scientist Peter Nordin at
RoboBusiness Conference in May 2007.
Brainstorm automatically adapts to onboard sensors and actuators, immediately builds a
model of any robot on which it is installed, and automatically writes control programs for the
robot’s movements. It can then explore and model its environment. Through simulated
interaction using these models, it solves problems and develops new behavior using
“imagination.” Once it has “learned” to do something, it can use its imagination to adapt its
behavior to a wide range of circumstances.
A methodology known as genetic programming (GP) is “the trick” that makes it all possible.
GP is an automated programming methodology inspired by natural evolution that is used to
evolve computer programs. Evolving computer programs means the logic developed by the
system can be anything that can be expressed by a computer program. That basically means
anything. Robots need descriptions of things they are supposed to do and they figure out how
to do them. GP itself is not an approach exclusive to robotic behavior. It has been applied to a
variety of problems, some already yielding commercial successes. An example well-known to
scientists in the field was the development of invention machines that had created two new
patentable inventions by 2002. The potential for “thinking robots” goes well beyond
developing their own actions.
The system is constructed using components and the learning / adaptive mechanisms can
be turned on and off. This provides a broad range of choices to satisfy requirements. It can for
example, be used for rapid development of control systems that cannot be modified after
testing is complete or the learning adaptive system can remain on during use allowing the
robot to continue to evolve as it gains real-life experience. The level of learning and
adaptation can be adjusted to requirements. It can be used to build robot software from the
ground up fulfilling all requirements or an add-on to an existing system that provides
learning and adaptive behavior. Although product development time can be significantly
shortened and less costly, it will still follow a familiar pattern. Product developers need to
define their product requirements and engineers will make decisions about the best
configurations and settings.
Some of the engineering work will be dramatically transformed however. Rather than
working on writing millions of lines of programming, engineers will focus on the best
descriptions of desired behavior and testing. iRobis can already provide a range of
predefined basic behaviors for mobile robots and plans to expand its behavioral library. The
fact that the system builds its own programs creates an opportunity for rapid research and
development through experimentation. Robots and environments can be constructed in
simulation where the behavioral programming will be automatically constructed. This allows
a great deal of experimentation on design to be carried out rapidly before committing to
construction of physical prototypes, allowing overall advancements to proceed at a faster
pace.
The research model for Brainstorm was developed in the 1990s at Chalmers University of
Technology in Gothenburg, Sweden. Dr. Nordin’s research there focused on self-developing
computer programs and developed a new architecture for intelligent robotics using GP. The
architecture was first presented in 1999 at The Fourth International Symposium on Artificial
Life and Robotics in Oita, Japan. (“An Evolutionary Architecture for a Humanoid Robot”) The
1999 presentation already included demonstration of robots that learned balance, human
gait, practical use of bifocal vision, navigation, audio orientation, hand-eye coordination,
and object manipulation. The 5 year long Humanoid Project at Chalmers furthered the work
with creation of hundreds of working robots of many types and sizes.
iRobis has been working on the commercial version under a “dual use” contract with the
Swedish Defense Department. Results are intended to benefit both military and non-military
technology. The announcement does not mean that the software is available to download for
a free 30-day trial. In the short-term, iRobis expects to work directly with early commercial
adopters and researchers to create prototypes with previously unseen levels of intelligent
autonomous behavior and to prove the value of the system for rapid development and
advanced experimentation. The search is on for the most valuable partners.
Institute of Robotics in Scandinavia AB (iRobis)
Vasaplatsen 2
41134 Gothenburg
Sweden
irobis.com
Contact:
Roger F. Gay
VP Business Development
Fiskarnasgata 161 v2
13662 Haninge
Sweden
If Intel makes digital CPUs and TI makes Digital Signal Processing (DSP for cellphones, etc.),
who will make neural networks for intelligent robots? Looks like it could be HP.
IBM Seeks to Build the Computer of the Future Based on Insights from the BrainIBM
Awarded DARPA Funding for Cognitive Computing Collaboration SAN JOSE, Calif. - 20 Nov
2008: In an unprecedented undertaking, IBM Research and five leading universities are
partnering to create computing systems that are expected to simulate and emulate the brain’
s abilities for sensation, perception, action, interaction and cognition while rivaling its low
power consumption and compact size. The digital data explosion shows no signs of slowing
down -- according to analyst firm IDC, the amount of digital data is growing at a mind-
boggling 60 percent each year, giving businesses access to incredible new streams of
information. But without the ability to monitor, analyze and react to this information in real-
time, the majority of its value may be lost. Until the data is captured and analyzed, decisions
or actions may be delayed. Cognitive computing offers the promise of systems that can
integrate and analyze vast amounts of data from many sources in the blink of an eye,
allowing businesses or individuals to make rapid decisions in time to have a significant
impact. For example, bankers must make split-second decisions based on constantly
changing data that flows at an ever-dizzying rate. And in the business of monitoring the
world’s water supply, a network of sensors and actuators constantly records and reports
metrics such as temperature, pressure, wave height, acoustics and ocean tide. In either
case, making sense of all that input would be a Herculean task for one person, or even for
100. A cognitive computer, acting as a “global brain,” could quickly and accurately put
together the disparate pieces of this complex puzzle and help people make good decisions
rapidly.By seeking inspiration from the structure, dynamics, function, and behavior of the
brain, the IBM-led cognitive computing research team aims to break the conventional
programmable machine paradigm. Ultimately, the team hopes to rival the brain’s low power
consumption and small size by using nanoscale devices for synapses and neurons. This
technology stands to bring about entirely new computing architectures and programming
paradigms. The end goal: ubiquitously deployed computers imbued with a new intelligence
that can integrate information from a variety of sensors and sources, deal with ambiguity,
respond in a context-dependent way, learn over time and carry out pattern recognition to
solve difficult problems based on perception, action and cognition in complex, real-world
environments. IBM and its collaborators have been awarded $4.9 million in funding from the
Defense Advanced Research Projects Agency (DARPA) for the first phase of DARPA’s Systems
of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) initiative. IBM’s proposal,
“Cognitive Computing via Synaptronics and Supercomputing (C2S2),” outlines
groundbreaking research over the next nine months in areas including synaptronics,
material science, neuromorphic circuitry, supercomputing simulations and virtual
environments. Initial research will focus on demonstrating nanoscale, low power synapse-
like devices and on uncovering the functional microcircuits of the brain. The long-term
mission of C2S2 is to demonstrate low-power, compact cognitive computers that approach
mammalian-scale intelligence. Exploratory research is in the fabric of IBM’s DNA,” said
Josephine Cheng, IBM Fellow and vice president of IBM’s Almaden Research Center in San
Jose. “We believe that our cognitive computing initiative will help shape the future of
computing in a significant way, bringing to bear new technologies that we haven’t even
begun to imagine. The initiative underscores IBM’s capabilities in bold, exploratory research
and interest in powerful collaborations to understand the way the world works.”IBM has
assembled a multi-dimensional, integrated world-class team of researchers and
collaborators led by Dr. Dharmendra Modha, manager of IBM’s cognitive computing
initiative, to take on the challenge including Stanford University (Professors Kwabena
Boahen, H.-S. Philip Wong, Brian Wandell), University of Wisconsin-Madison (Professor
Gulio Tononi), Cornell University (Professor Rajit Manohar), Columbia University Medical
Center (Professor Stefano Fusi) and University of California- Merced (Professor Christopher
Kello). IBM Researchers include Dr. Stuart Parkin, Dr. Chung Lam, Dr. Bulent Kurdi, Dr. J.
Campbell Scott, Dr. Paul Maglio, Dr. Simone Raoux, Dr. Rajagopal Ananthanarayanan, Dr.
Raghav Singh, and Dr. Bipin Rajendran. Recently, the IBM cognitive computing team
demonstrated the near-real-time simulation at a scale of a small mammal brain using
cognitive computing algorithms with the power of IBM’s BlueGene supercomputer. With this
simulation capability, the researchers are experimenting with various mathematical
hypotheses of brain function and structure as they work toward discovering the brain’s core
computational micro and macro circuits. In the past, the field of artificial intelligence
research has focused on individual aspects of engineering intelligent machines. Cognitive
computing, on the cutting edge of this line of research, seeks to engineer holistic intelligent
machines that neatly tie together all of the pieces. IBM’s cognitive computing initiative was
born out its 2006 Almaden Institute, which annually brings together top minds to address
fundamental challenges at the very edge of science and technology. IBM has a rich history
in the area of artificial intelligence research going all the way back to 1956 when IBM
performed the world’s first large-scale (512 neuron) cortical simulation.For more information
about IBM Research, please visit www.ibm.com/research or the IBM Research blog at: http:
//ibmresearchnews.blogspot.com/.
As we look for computers and robots to approach the capabilities of the human mind, we find
that the density of storage in computers is already greater than the density of the human
brain except for three factors. Computer chips consume more power than the human brain.
Computer chips are two-dimensional and the human brain is three-dimensional. The
human brain uses some elements like a memristor and current computers do not use
memristors as yet. As a result, progress is punctuated by three things, lower power, three
dimensionality and the use of memristors. Genetic Programming must also be developed
further so that the robot may learn rather than being programmed.
Electronic two-terminal bistable graphitic memories
Transistors are the basis for electronic switching and memory devices as they exhibit
extreme reliabilities with on/off ratios of 10,000-100,000, and billions of these three-
terminal devices can be fabricated on single planar substrates. On the other hand, two-
terminal devices coupled with a nonlinear current-voltage response can be considered as
alternatives provided they have large and reliable on/off ratios and that they can be
fabricated on a large scale using conventional or easily accessible methods. Here, we report
that two-terminal devices consisting of discontinuous 5-10 nm thin films of graphitic sheets
grown by chemical vapor deposition on either nanowires or atop planar silicon oxide exhibit
enormous and sharp room-temperature bistable current-voltage behavior possessing stable,
rewritable, non-volatile and non-destructive read memories with on/off ratios of up to
10,000,000 and switching times of up to 1 microsecond (tested limit). A
nanoelectromechanical mechanism is proposed for the unusually pronounced switching
behavior in the devices.
- it would increase the amount of storage in a two-dimensional array by a factor of five.
- the new switches can be controlled by two terminals instead of three, as in current chips.
This will allow for practical three dimensional memory layers.
- being essentially a mechanical device, such chips will consume virtually no power when
storing memory
- On/off power ratio of one million to one instead of ten to one for phase change memory
[higher is better]
- James Tour said the new switches are also fast; in fact, they react faster than his lab's
current testing systems can measure.
- they're robust. "We've tested it in the lab 20,000 times with no degradation," said Tour. "Its
lifetime is going to be huge, much better than flash memory."
- Typically, graphene is very hard to think about fabricating commercially," he said, "but this
can be done very easily by deposition.
Taken from Nature: Materials.
3-D memristor chip debutsR. Colin JohnsonEE Times (11/26/2008 10:36 AM EST)
PORTLAND, Ore. — Memristors technology got a boost recently from Hewlett-Packard Labs,
which described the first 3-D memristor chip at a conference in Berkeley, Calif. The
Memristor and Memristive Systems Symposium was co-sponsored by the University of
California, the Semiconductor Industry Association and the National Science Foundation.
HP Labs (Palo Alto, Calif.) provided details of a prototype chip designed by HP researcher
Qiangfei Xia that stacked memristor crossbar memory cells on top of a CMOS logic chip. "Xia
used imprint lithography to add a memristor crossbar on top of a CMOS logic circuit," said HP
Labs Fellow Stan Williams, inventor of HP's memristive memory technology. "He has built an
integrated hybrid circuit with both transistors and memristors." Williams and HP colleague
Greg Snider previously proposed an FPGA in which configuration bits were located above
CMOS transistors in a memristor crossbar. Memristor crossbars include two titanium dioxide
layers between two perpendicular arrays of metal lines. One layer of titanium oxide is doped
with oxygen vacancies, making it a semiconductor. The adjacent layer is undoped, leaving it
in its natural state as an insulator. When a crossbar junction is addressed by
simultaneously applying a voltage to one crossbar line on the top and bottom layers, oxygen
vacancies drift from the doped to the undoped layer. This causes it to begin conducting,
turning "on" the memory bit. The bit can again be turned "off" by changing the current
direction, whereupon oxygen vacancies migrate back into the doped layer. According to
Williams, HP Labs' memristor-based FPGA demonstrates that a CMOS fab can make
integrated memristor/transistor circuits in three dimensions. Also at the symposium, Snider
unveiled a design that used memristors in their analog mode as synapses in a neural
computing architecture. Memristor crossbars are the only technology that is dense enough
to simulate the human brain, Snider claimed, adding that the HP Labs crossbars are ten
times denser than synapses in the human cortex. By stacking crossbars on a CMOS logic
chip, variable resistance could mimic the learning functions of synapses in neural
networks. HP Labs and Boston University were recently awarded a contract by the Defense
Advanced Research Projects Agency to build the first artificial neural network based on
memristors. Also at the conference, Massimiliano Di Ventra of the University of California at
San Diego described how memristors can explain biological learning in amoebas. Amoebas
learn to change their behavior in a manner that can be explained by an LC circuit and a
memristor. Di Ventra also presented evidence that microscopic memristive elements are
present in unicellular as well as multicellular organisms.
iRobis Announces Complete Cognitive Software System for Robots
IT, New Media & Software
Press release from: iRobis
(openPR) - November 27, 2008 - Institute of Robotics in Scandinavia (iRobis) has announced
that the world’s first “complete cognitive software system for robotics” is ready for application.
The system turns robots into self-developing, adaptive, problem-solving, “thinking”
machines.
VP for Business Development Roger F. Gay made the announcement November 19th at the
RoboDevelopment Conference & Expo in Santa Clara, California. Development of the software
system, called Brainstorm™, was previously discussed by lead scientist Peter Nordin at
RoboBusiness Conference in May 2007.
Brainstorm automatically adapts to onboard sensors and actuators, immediately builds a
model of any robot on which it is installed, and automatically writes control programs for the
robot’s movements. It can then explore and model its environment. Through simulated
interaction using these models, it solves problems and develops new behavior using
“imagination.” Once it has “learned” to do something, it can use its imagination to adapt its
behavior to a wide range of circumstances.
A methodology known as genetic programming (GP) is “the trick” that makes it all possible.
GP is an automated programming methodology inspired by natural evolution that is used to
evolve computer programs. Evolving computer programs means the logic developed by the
system can be anything that can be expressed by a computer program. That basically means
anything. Robots need descriptions of things they are supposed to do and they figure out how
to do them. GP itself is not an approach exclusive to robotic behavior. It has been applied to a
variety of problems, some already yielding commercial successes. An example well-known to
scientists in the field was the development of invention machines that had created two new
patentable inventions by 2002. The potential for “thinking robots” goes well beyond
developing their own actions.
The system is constructed using components and the learning / adaptive mechanisms can
be turned on and off. This provides a broad range of choices to satisfy requirements. It can for
example, be used for rapid development of control systems that cannot be modified after
testing is complete or the learning adaptive system can remain on during use allowing the
robot to continue to evolve as it gains real-life experience. The level of learning and
adaptation can be adjusted to requirements. It can be used to build robot software from the
ground up fulfilling all requirements or an add-on to an existing system that provides
learning and adaptive behavior. Although product development time can be significantly
shortened and less costly, it will still follow a familiar pattern. Product developers need to
define their product requirements and engineers will make decisions about the best
configurations and settings.
Some of the engineering work will be dramatically transformed however. Rather than
working on writing millions of lines of programming, engineers will focus on the best
descriptions of desired behavior and testing. iRobis can already provide a range of
predefined basic behaviors for mobile robots and plans to expand its behavioral library. The
fact that the system builds its own programs creates an opportunity for rapid research and
development through experimentation. Robots and environments can be constructed in
simulation where the behavioral programming will be automatically constructed. This allows
a great deal of experimentation on design to be carried out rapidly before committing to
construction of physical prototypes, allowing overall advancements to proceed at a faster
pace.
The research model for Brainstorm was developed in the 1990s at Chalmers University of
Technology in Gothenburg, Sweden. Dr. Nordin’s research there focused on self-developing
computer programs and developed a new architecture for intelligent robotics using GP. The
architecture was first presented in 1999 at The Fourth International Symposium on Artificial
Life and Robotics in Oita, Japan. (“An Evolutionary Architecture for a Humanoid Robot”) The
1999 presentation already included demonstration of robots that learned balance, human
gait, practical use of bifocal vision, navigation, audio orientation, hand-eye coordination,
and object manipulation. The 5 year long Humanoid Project at Chalmers furthered the work
with creation of hundreds of working robots of many types and sizes.
iRobis has been working on the commercial version under a “dual use” contract with the
Swedish Defense Department. Results are intended to benefit both military and non-military
technology. The announcement does not mean that the software is available to download for
a free 30-day trial. In the short-term, iRobis expects to work directly with early commercial
adopters and researchers to create prototypes with previously unseen levels of intelligent
autonomous behavior and to prove the value of the system for rapid development and
advanced experimentation. The search is on for the most valuable partners.
Institute of Robotics in Scandinavia AB (iRobis)
Vasaplatsen 2
41134 Gothenburg
Sweden
irobis.com
Contact:
Roger F. Gay
VP Business Development
Fiskarnasgata 161 v2
13662 Haninge
Sweden
If Intel makes digital CPUs and TI makes Digital Signal Processing (DSP for cellphones, etc.),
who will make neural networks for intelligent robots? Looks like it could be HP.
IBM Seeks to Build the Computer of the Future Based on Insights from the BrainIBM
Awarded DARPA Funding for Cognitive Computing Collaboration SAN JOSE, Calif. - 20 Nov
2008: In an unprecedented undertaking, IBM Research and five leading universities are
partnering to create computing systems that are expected to simulate and emulate the brain’
s abilities for sensation, perception, action, interaction and cognition while rivaling its low
power consumption and compact size. The digital data explosion shows no signs of slowing
down -- according to analyst firm IDC, the amount of digital data is growing at a mind-
boggling 60 percent each year, giving businesses access to incredible new streams of
information. But without the ability to monitor, analyze and react to this information in real-
time, the majority of its value may be lost. Until the data is captured and analyzed, decisions
or actions may be delayed. Cognitive computing offers the promise of systems that can
integrate and analyze vast amounts of data from many sources in the blink of an eye,
allowing businesses or individuals to make rapid decisions in time to have a significant
impact. For example, bankers must make split-second decisions based on constantly
changing data that flows at an ever-dizzying rate. And in the business of monitoring the
world’s water supply, a network of sensors and actuators constantly records and reports
metrics such as temperature, pressure, wave height, acoustics and ocean tide. In either
case, making sense of all that input would be a Herculean task for one person, or even for
100. A cognitive computer, acting as a “global brain,” could quickly and accurately put
together the disparate pieces of this complex puzzle and help people make good decisions
rapidly.By seeking inspiration from the structure, dynamics, function, and behavior of the
brain, the IBM-led cognitive computing research team aims to break the conventional
programmable machine paradigm. Ultimately, the team hopes to rival the brain’s low power
consumption and small size by using nanoscale devices for synapses and neurons. This
technology stands to bring about entirely new computing architectures and programming
paradigms. The end goal: ubiquitously deployed computers imbued with a new intelligence
that can integrate information from a variety of sensors and sources, deal with ambiguity,
respond in a context-dependent way, learn over time and carry out pattern recognition to
solve difficult problems based on perception, action and cognition in complex, real-world
environments. IBM and its collaborators have been awarded $4.9 million in funding from the
Defense Advanced Research Projects Agency (DARPA) for the first phase of DARPA’s Systems
of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) initiative. IBM’s proposal,
“Cognitive Computing via Synaptronics and Supercomputing (C2S2),” outlines
groundbreaking research over the next nine months in areas including synaptronics,
material science, neuromorphic circuitry, supercomputing simulations and virtual
environments. Initial research will focus on demonstrating nanoscale, low power synapse-
like devices and on uncovering the functional microcircuits of the brain. The long-term
mission of C2S2 is to demonstrate low-power, compact cognitive computers that approach
mammalian-scale intelligence. Exploratory research is in the fabric of IBM’s DNA,” said
Josephine Cheng, IBM Fellow and vice president of IBM’s Almaden Research Center in San
Jose. “We believe that our cognitive computing initiative will help shape the future of
computing in a significant way, bringing to bear new technologies that we haven’t even
begun to imagine. The initiative underscores IBM’s capabilities in bold, exploratory research
and interest in powerful collaborations to understand the way the world works.”IBM has
assembled a multi-dimensional, integrated world-class team of researchers and
collaborators led by Dr. Dharmendra Modha, manager of IBM’s cognitive computing
initiative, to take on the challenge including Stanford University (Professors Kwabena
Boahen, H.-S. Philip Wong, Brian Wandell), University of Wisconsin-Madison (Professor
Gulio Tononi), Cornell University (Professor Rajit Manohar), Columbia University Medical
Center (Professor Stefano Fusi) and University of California- Merced (Professor Christopher
Kello). IBM Researchers include Dr. Stuart Parkin, Dr. Chung Lam, Dr. Bulent Kurdi, Dr. J.
Campbell Scott, Dr. Paul Maglio, Dr. Simone Raoux, Dr. Rajagopal Ananthanarayanan, Dr.
Raghav Singh, and Dr. Bipin Rajendran. Recently, the IBM cognitive computing team
demonstrated the near-real-time simulation at a scale of a small mammal brain using
cognitive computing algorithms with the power of IBM’s BlueGene supercomputer. With this
simulation capability, the researchers are experimenting with various mathematical
hypotheses of brain function and structure as they work toward discovering the brain’s core
computational micro and macro circuits. In the past, the field of artificial intelligence
research has focused on individual aspects of engineering intelligent machines. Cognitive
computing, on the cutting edge of this line of research, seeks to engineer holistic intelligent
machines that neatly tie together all of the pieces. IBM’s cognitive computing initiative was
born out its 2006 Almaden Institute, which annually brings together top minds to address
fundamental challenges at the very edge of science and technology. IBM has a rich history
in the area of artificial intelligence research going all the way back to 1956 when IBM
performed the world’s first large-scale (512 neuron) cortical simulation.For more information
about IBM Research, please visit www.ibm.com/research or the IBM Research blog at: http:
//ibmresearchnews.blogspot.com/.
 | |  | |  | |  |