Neurobots: Robots Controlled by Brain Simulations

Researchers have been developing robots that are powered by better artificial brains. They have recently created a computer neural simulation consisting of 6,700 neurons with approximately 1.3 million synaptic connections. This technology builds on previous work in the field of neurorobotics. The robot they used for this experiment is shown on the left. It is equipped with a CCD video camera. The camera has IR sensors to avoid obstacles and an RF transmitter to process objects visually. The emulation attempts to model aspects of the mind that researchers believe to be important for information processing. The goal is to allow the robot to act in a manner similar to an actual animal. This gives insight into the functioning of the brain and how it encodes for behavior.

The scientists claim to have chosen three different neurotransmitters subsystems to model. These systems include dopamine, acetylcholine and serotonin. Dopaminergic, cholinergic and serotonergic cell bodies are found in the VTA, basal forebrain and raphe nucleus respectively. Some of these cells are located deep within the brain and they project their axons like a branching tree to numerous other regions. The synaptic junctions allow cross-talk between areas using these discrete neurotransmitters as messengers.

The VTA for instance (see picture on left), makes connections to places like the nucleus accumbens and prefrontal cortex. The neural facsimile they carried out emulates 100 neurons for each of these three cell body regions. Their simulation also contains brain cells devoted to processing images from the CCD camera. In addition, they have areas to encode for find/flee behavior and good/bad judgements (100 neurons for each).

So this is really an extremely simplified replica that is merely meant to represent some very basic ways that the researchers believe how the mind works. Obviously 6,700 neurons is much less complex than even a fly's brain. They ignored a considerable amount of neuronal function. So the behavioral output of this robotic device is definitely limited in scope compared to more complicated organisms.

The researchers appear to give an overly simplistic explanation for how specific neurotransmitter systems function. So I'm not necessarily convinced of the utility in labeling the 100 neuron subpopulations as "dopaminergic" or "serotonergic". Their brain is really only a crude simulacrum and assigning these labels may not be particularly relevant.

Dopamine appears to be important for “wanting”, that is, the motivation process in acquiring an object [13]. Dopamine, which is found throughout the central nervous system, is produced in the ventral tegmental area. A recent proposal ties the prediction error to wanting by suggesting that incentive salience is the expected future reward that maps actions to rewards [14].

There are a lot of nuances to dopamine's role in motivating an organisms to action and I'm not confident that the authors do it justice in their paper. Since this is done via computer, it does allow the scientists to do temporary lesions to see the resulting affect on robotic performances. They can basically turn off the functioning of specific neurotransmitter subsystems selectively. I'd be wary, though, of linking the behavioral changes that they witnessed to a real animal's dopaminergic system.

When CARL-1’s dopaminergic system was intact, it approached stimuli that were predictive of positive value, and ignored neutral stimuli. When CARL-1’s VTA was lesioned, the number of Find responses, which signify “wanting”, significantly decreased. Instead of approaching these positive-value stimuli, CARL-1 treated green objects as neutral stimuli.

They go on to talk about the other neuromodulatory systems and how adjusting them altered the functioning of the neurobot.

In our experiments with CARL-1, we showed that serotonergic neuromodulation arising from a simulated Raphe nucleus was needed to respond appropriately to threatening stimuli. When CARL-1’s serotonergic system was intact, it moved away from threatening stimuli, and ignored neutral stimuli.

The interpretations they make seem somewhat facile to me. However, I do think the interesting aspect of neurorobotics is that it allows the researchers to test an extraordinary range of different hypotheses. There is a lot of potential in scaling up the neuron count to enable a more extensive range of robot routines.

You can see some videos of this robot here. A recent paper on the topic is here (PDF). The researcher also gave a video presentation about it (see here, requires flash).

Read more...

Neural Interface

I found some more information about the HIVE project. A presentation was given November of last year discussing the potential of computer controlled brain stimulation (see PDF). The researchers definitely appear to have an eye towards some more futurist speculative uses of the technology.

10 Mapping our brains to computers (the singularity)
9 Jacking in (invasive interaction)
8 Non-invasive Brain 2 Machine + Machine 2 Brain interaction
7 Immersion (HMD/CAVE + haptics + ...) (also MR/AR) using natural senses

There is also a new article in AlphaGalileo about it as well. Here's an excerpt (translated from spanish);

One case of possible application that this (technology) poses to the future researcher Pablo de Olavide is in the treatment of some types of deafness. In this line, the device developed could be applied within a few years to develop a stimulus pattern that simulates human speech or sound, for people who can not hear through the ear, can get the information directly into your brain. "In these cases, the inner ear that fails, not the brain, so the device could be applied to stimulate the brain related to hearing," concludes the researcher.

Beaming sensory experiences into the brain could be helpful for those with certain disabilities. Scientists have also been utilizing brain research in order to facilitate the development of more engrossing and authentic virtual realities. Due to increases in GPU power, virtual environments will likely become more representative of actual real world circumstances as time goes forward. More theoretical technology might eventually enable computer generated sensations to be directly transmitted into the minds of normal people. I think some intriguing things could happen as this field matures. Being able to generate any sort of qualia on command via a digital program is basically the ultimate end point. Coupling that ability with more exact methods of fine tuning how the brain actually perceives qualia could usher in a transformative shift in consciousness.

Read more...

Graphics Processing Unit (GPU) for Brain Research

Graphics Processing Units (GPU) are commonly used to power video game software. However, they are also finding use for a more diverse array of scientific research as well. A GPU conference has recently taken place discussing some of the applications of this technology (see PDF 8.1 MB). Here are a few excerpts about the GPU brain projects. One deals with the connectome (circuit diagram).

Determining the detailed connections in brain circuits is a fundamental unsolved problem in neuroscience. Understanding this circuitry will enable brain scientists to confirm or refute existing models, develop new ones, and come closer to an understanding of how the brain works. Prof. Jeff Lichtman and Center for Brain Science (CBS) at Harvard launched the Connectome Project three years ago to determine the complete, detailed wiring diagrams of neural circuits from sequential high-resolution images of the central nervous system using electron microscopy (EM). These high-resolution, large-scale EM datasets pose very challenging computational problems for 3D segmentation and visualization in terms of developing suitable algorithms, coping with the ever-increasing data sizes, and maintaining interactive performance.

Visual recognition software is another area that this tech could speed up performance.

Nicolas Pinto is a second-year PhD Student in Computational Neuroscience at MIT. He is currently a member of the DiCarlo Lab and the Sinha Lab at MIT, and the Visual Neuroscience Group at Harvard. His research interests lie at the intersection of Brain and Computer Sciences. The overarching goal of his research is to dramatically accelerate the development of computational theories of how the visual cortex accomplishes object recognition. In addition to advancing our understanding of how the brain works by generating new experimentally testable hypotheses, this approach also holds great promise for the development of new artificial vision systems. A key innovation in his work is the ability to leverage the computational power of disruptive technologies like NVIDIA’s GPUs to provide new insights into this fundamental problem.

A Harvard researcher has recently talked about how these new methods will enable us to answer many of the big questions. From the Big Bang (and even before then) to the evolution of humans, computing power will truly help us understand almost any question imaginable. Better supercomputers may lead to complete and detailed simulations of living tissue. Researchers are developing multi-scale modeling from bio-molecules to organs (see PDF). With the help of these virtual models we will essentially be able to reprogram our own brain and body matter. Are we headed toward ageless bodies and superhappy minds? Only time will tell what new avenues this kind of processing power will open up.

See also GPU-Based Petascale Visual Computing for Analysis of Neural Circuitry (PDF).

Read more...

Electron Microscope to Image Living Cells

I have previously mentioned about increasing the resolving power of light based microscopes in order to better image living tissue. Electron microscopes have an even greater ability to view finer details. The only problem is that focused electron beams can easily damage living cells. Now scientists are using the properties of quantum mechanics in order to develop electron microscopy that would be able to create pictures of that type of tissue without destroying it. The researchers believe that eventually this will allow them to achieve a resolution of several nanometers. At this level of detail they could view individual molecules inside of cells. Obviously being able to view bio-molecular interactions within a neuron could have implications for improved understanding of their functioning. You can read more about this interesting advance in the press release.

Read more...

Brain-Computer Interface and the Wireless Neurosociety

Investigators have been creating superior wireless brain-computer interfaces (BCI). Being able to shed wires has the promise of enhancing the usability of these devices for those with disabilities. As time goes forward we may increasingly become a wireless neurosociety. This has the potential to irrevocably transform how we relate to others and interact with the environment around us. New tools may enhance our ability to manipulate the world and allow an unprecedented new means of communication with both computers and people.

Some scientists are additionally working on synthetic telepathy. This research basically entails capturing EEG brain readings that are the neural correlates of our inner monologue. These signals would then be translated by a computer into a voice synthesizer. This would allow a person to correspond with someone else without even opening their mouth. They would merely have to "think" about what they wanted to say and then that could be wirelessly beamed into an ear phone on another person. Theoretically new neuromodulation methods may also be used to artificially generate voices without the need for an ear phone. Brain implants (or perhaps non-invasive ultrasonic neuromodulation using an external device) that stimulated subpopulations of neurons associated with the perception of hearing might allow the creation of hallucinatory sounds. You would be able to perceive someone else talking clearly in your head. This could be useful in the military because this type of communication would generate no audible noise whatsoever. It could allow a two-way dialogue between soldiers using waves on the electromagnetic spectrum.

Researchers are also developing smart homes that could be controlled by brain computer interfaces. Imagine being able to turn on your television, brighten lights or open doors solely with the power of your own mind. A thought reading helmet that could allow people to fly an airplane with their brain power is in the works as well. So it seems possible that a single sophisticated BCI may enable a person to exert control over their house, their car and communicate with others telepathically. Also, why type on a keyboard when you can just employ thoughts to disseminate information to your computer? All of your inner monologue could be continuously and automatically written down for you on a word software program.

Brain computer interfaces of the future may both decipher brain signals and manipulate them as well. Better deep brain stimulation implants are already finding increased utilization among people with specific disorders. Complex computer controlled brain stimulation may increasingly become the norm. However, there may be many issues that come up with regards to this new technology. These devices have the potential of being hacked by outsiders. Researchers have begun to consider the ramifications of these types of privacy issues for DBS implants. The fact that they now have wireless inputs means that they can be maliciously hijacked or spied upon by another person.

In a future society, some people may adopt more drastic types of implants for themselves. Being able to access information from the web and have it beamed directly into your head could be a tremendous boon for learning. The rate at which people acquire and manipulate data would increase at a tremendous rate. However, these types of direct connections to the net also bring up the same issues of privacy as with the less sophisticated neural gear. More complicated brain apparatuses might be susceptible to contracting some sort of virus that could radically alter the functioning of the appliance.

Imagine if you had a brain implant to improve memory and it stored a copious amount of information. A virus transferred by a wireless signal could possibly rewrite specific types of past recollections, thus altering everything about your previously remembered life history. Or perhaps a hacker or rogue AI program might adjust a person's behavior in a specific way. Maybe they could gain a top level control over someone else and turn them into some sort of botnet drone. Also other people might be able to gain direct access to the private introspection of another person. A wireless brain computer interface that recorded thoughts could potentially be spied upon, much like computers connected to the internet can be today by spyware. This may give a person or government insight into what someone was contemplating ahead of time. Certain countries might gravitate towards this possibility of better controlling or understanding their own citizens. Specific companies might also want the chance to broadcast advertisements wirelessly to a person's brain or gain access to what sorts of products the person would want to buy. If you can hear these messages within your head, how do you prevent your mind from being overrun by spam?

Some people may actually choose to allow others to eavesdrop on their own cognitive processes. This would be analogous to how many people use twitter to broadcast some of their succinct ruminations to whoever will listen. You could potentially selectively choose who you want to overhear your thoughts and block others from access. Will people in the future use neuromodulatory techniques to shed their inhibitions and allow a totally open society? A sousveillance where anyone can listen in on anyone else's internal monologue? Maybe a minority of people would even prefer to have outsiders control their behavior to a certain extent with targeted rewarding brain stimulation or another type of computer controlled mind manipulation.

Perhaps in the future we will also be able to send and receive nuanced emotions along with thoughts. A brain implant could acquire signals and then stimulate brain regions associated with certain feelings. This would be the next step in human evolution and would supercede regions of the mind currently involved in empathetic awareness. We would finally be able to truly feel others joy and pain directly instead of the roundabout way we currently do. I think there are a few other interesting question pertaining to this for future scientists to figure out. A main one being; how much of our consciousness or emotions can we transmit by using electromagnetic radiation?

Many of these things are now highly speculative. Brain-computer interfaces still have a long ways to go before they would have some of these capabilities. The adoption of any said technology may also depend on how easy or practical it is to use. The actual utilization of BCI's rests on the vagaries of future human desires and not what may theoretically be possible. However, there is definitely a lot of interest in improving this sort of technology. BCI's are already entering the market to enable people to play video games with their minds, for instance. So there are a number of interesting future scenarios that could crop up as time goes forward. A wireless neurosociety could potentially be a significant change from what people are currently accustomed to.

Read more...

Transformative Brain Projects

There are a couple new "transformative" brain projects that will begin soon. Neuroscience grants have been awarded to two people in order to further our understanding of how the mind functions. Here's an excerpt about the first project (connectome);

Mitra and colleagues, including Professor Harvey Karten of the University of California, San Diego, will use their “transformative” grant to produce the first brain-wide circuit diagram for the mouse, and using this as reference, attempt to determine alterations in the corresponding circuits of mouse models of neuropsychiatric disorders.

Below is an excerpt discussing the second project;

Josh Dubnau’s “transformative” project addresses an important gap in knowledge: about how this fundamental step in the conversion of genetic information -- its “translation” from RNA to protein -- is regulated in neurons, the ubiquitous cells of the brain whose dense web of connections underlie its capacity to perform sophisticated functions such as forming and storing memories.

There has also recently been talk about creating a complete connectome wiring diagram of the human brain. The NIH has aimed to do this within 5 years. A neuroscience blogger has shown skepticism about this. He thinks that the brain is far too complex and it will actually take much longer to get this diagram. I would partially disagree with his points. I believe it is important not to take an overly linear view of progress. Yes it seems like a daunting task. However, researchers are continuously creating better tools in order to acquire this type of data faster. I wouldn't argue that it will necessarily happen within 5 years, but I think the speed at which it occurs will be suprising.

Ray Kurzweil talks a lot about certain accelerating trends or (s-curves). Certain technologies don't progress in a linear rate, but much faster. In his book, Kurzweil gives an example;

"When the human-genome scan got under way in 1990 critics pointed out that given the speed with which the genome could then be scanned it would take thousands of years to finish the project. Yet the fifteen-year was completed slightly ahead of schedule, with a first draft in 2003.

Now the amount of people who have had their genome sequenced is probably going to increase at an exponential rate over the course of the next several years. It won't be long before everyone who wants to have their genome sequenced will be able to have it done. This has been happening because new tools have allowed for faster and cheaper sequencing of DNA. A main problem with Kurzweil is that he takes his accelerating trends analysis too far and tries to apply it to things where it doesn't work. Also these accelerating trends do end eventually, which Kurzweil doesn't spend enough time discussing. So while some of the points he makes are good, he is not necessarily the most reliable source. Overall, though, I think it is important to have a broader understanding of specific trends that exist in a variety of different fields. Many scientists/neuroscientists may have an overly narrow focus of what they study and it is difficult for any one person to keep abreast of developments in other fields. They may not be totally aware of scientific progress in unrelated disciplines, so they might underestimate what could be possible to do with technology and how fast it will occur.

Read more...

Neurosystems for National Security

I came across a research program at the Mind Research Network. It's called the Neurosystems for National Security and looks like it deals with the applications of brain technology towards improving the functioning of military personnel. Here's an excerpt;

MRN possesses the unique ability to utilize and combine functional imaging and brain scanning techniques (fMRI, MEG, EEG), computer modeling and simulation, cortical brain stimulation and genetics to investigate how the brain functions and how it can be made to function better for the safety, security, and reliability of our military and national security interests.

One potential benefit involves helping military and national security personnel make better decisions under stress. Biological changes occur in the brain and body in response to stress. These stress responses are intended to serve adaptive functions, but can also have a negative influence on cognition and behavior. One of our goals is to develop methods and techniques to leverage and modulate stress to optimize decision making. The ability to better modulate stress in times of crisis would be invaluable to both the foot soldier under fire and the general commander making critical national security decisions.

Many militaries have shown interest in brain research (see neurowarfare report). There are obviously a lot of ethical issues to this type of stuff. Should we really be pushing to use brain tools to improve the functioning of soldiers? There may be less benign things like using transcranial magnetic stimulation to reduce cognitive deficits associated with fatigue. Obviously if you could increase the sum total of cognitive or creative capacity of personnel it could have a huge effect on how well the military performs. There may be other radical stuff that might crop up in the next ten years that could be more ethically dubious.

Understanding the underpinnings of the stress response could enable soldiers who are better able to cope with being in battle. I've mentioned about using neurotechnology to amplify feelings of empathy. However, there is also the converse. New tools of neuromodulation might possibly selectively reduce these feelings temporarily to enable better soldiers who are less concerned about killing others. The insular cortex is a region of the mind that is involved with feelings of disgust. Perhaps you would alter activity in this area with certain drugs or brain manipulation techniques in order to regulate how disgusted a person felt from their actions. The intensity of other negative feelings like fear also might be lessened in severity. Beta blockers have shown promise in weakening the experience of bothersome traumatic fearful memories for instance. Maybe they could adjust activity in the anterior cingulate cortex in order to blunt feelings of pain as well. I think even minor alterations in soldiers brain functioning can be problematic from an ethical standpoint. It could be used by the government to make soldiers more likely to stay in the military or follow orders.

Brain technology has other applications to national security issues. Researchers have shown interest in using TMS/tDCS for detecting and altering deceptive behavior. Will these new tools allow a person to cooperate more with authorities? Perhaps they could be used to make a prisoner more likely to tell the truth. As these devices become more refined, it may become easier to alter a person's behavior in a specific way. Ultrasonic neuromodulation might potentially be used to change activity in reward related brain regions for positive reinforcement. There is also the possibility of using these methods to non-invasively modulate areas of the mind associated with pain for torture. I think we will have to have robust defenses against these sorts of abuses by people in authority.

Deaths from conflict have been on the decline over the past 50 years. So I'm somewhat optimistic that new tools will be beneficial for humanity as opposed to making things worse. Perhaps people in the future will choose to modify their behavior in order to edit out warlike tendencies. These technologies should theoretically enable people to enhance feelings of being one and at peace with others. Humanity may eventually change their temperament to such a radical extent that almost no conflicts will occur. This would be an extreme shift in how the world operates. Regardless of what actually happens, there are many interesting issues with regards to neurotechnology that our society may increasingly have to grapple with as time goes forward.

See Mind Research Network Sponsors Lecture on Neurosystems for National Security.

Read more...

Neuron Imaging -Increasing Resolution Threefold

Scientists are continuously refining tools to better delineate the details of the brain's cellular workings. One method of capturing extremely small aspects of brain cells is by using two-photon microscopy. This type of device can pick up minutiae that other scanners like magnetic resonance can't. Now researchers have increased the resolution of this imaging modality by threefold. They have combined stimulated emission depletion microscopy with the two-photon microscopy in order to gain this ability. This new method will enable them to see things like dendritic spines on neurons with more clarity than ever before. This will lead to a better understanding of how the connections between neurons function and change over time.

With conventional microscope lenses there is something called the "diffraction limit". This limit means that light can not be focused to less than half of its wavelength. However, researchers have found many news ways of circumventing this limit. Stimulated emission depletion (STED) microscopy is one of the first methods that overcame this limit. Researchers are continuing to better focus light so as to characterize smaller cellular features. Recently scientists have been able to focus light to 20 times smaller than its normal wavelength. In the future, the researchers hope to shrink light so it is comparable to the size of an electron's wavelength. An electron is an elementary particle that is much smaller than a single atom. Electron microscopes use beams of electrons for imaging purposes and have extremely high resolving power. The only problem is that they can't be used on live tissue, while light based microscopes can.

The researchers who have combined the 2-photon with the STED microscopy believe that they can already increase the resolution another threefold. This would be done by exactly timing the pulse of the depletion light. I think there will be a push to incorporate newer technologies to allow for even better resolving power in the future. This will lead to an improved understanding of how the brain functions at a sub-cellular level.

Read more...

Neuromorphic AI Intelligent Text Recognition

DARPA has shown interest in creating artificially intelligent text reading systems recently. The air force research laboratory has been developing neuromorphic programs that are inspired by how the brain functions. They are creating software that can fill in missing portions of sentences (PDF) based on the context of the previous words.

Modern pattern recognition technology can perform accurately at its job when images are complete and easily observable. However when there is only a partial text image, a computer's accuracy pales in comparison to the human brain. The human brain is able to fill in details based on contextual relationships of surrounding words. Researchers have now been able to get a computer to make up missing details in sentences. They have mimicked aspects of the mind that are involved with visual processing in order to carry out this task. With a sufficiently advanced system this might allow a computer to understand text or speech that is garbled or partially missing.

There are a variety of different "levels" of the mind that people are trying to emulate. I've mention about the Blue Brain project that is attempting to replicate neurons and synaptic connections in software. That project may become even more detailed than that as time goes forward. There are also somewhat less detailed simulations that may not go into that much depth and use more simplified neuron configurations. At another end, there are people who are attempting behavioral modeling. Instead of emulating brain cells/synapses, they are basically looking at the overall modularity of the brain and how certain areas function together in order to enable specific states of cognition to exist.

The air force has been focusing on simulations that are in between the more detailed Blue Brain and the less complicated behavioral models. This level of detail is at the cortical column level. They mention there are approximately 10^8 minicolumns in the neocortex. There are about 100 brain cells in each minicolumn and simulating connectivity between them is easier than trying to account for individual neuronal connections. Each of the 32 axons from a single minicolumn extend to 32 other minicolumns. So in the neocortex, they estimated that there are 10^11 connections at this level compared to 10^14 connections at the neuron scale. They are currently modeling brain areas like the lateral geniculate nucleus (LGN) and the primary visual cortex (V1). As time goes forward they may be able to make replicas of the functioning of other brain regions that are important for processing information.

They have investigated several different models. One is a bayesian model of invariant pattern recognition (see left picture). A representation of the visual cortex has already been developed using this bayesian framework. That brain region is important for processing what a person sees. This design can allow for a neuromorphic AI to be able to identify items.


The bayesian model has previously been tested using the images on the left. The model had an approximately 50% recognition rate of these 32 by 32 pixel arrays. The company Numenta Inc. has been developing this software independently of the military. Numenta has been continuously refining the technology for more sophisticated visual identification under more varied conditions.

The scientists also researched a network of attractors. They discuss the "Ersatz Brain Project". This project is an effort to replicate aspects of how the mind functions by using nested networks of fixed point attractors. This is more of an algorithmic method of mimicking parts of the brain. They specifically picked the brain state in a box (BSB) algorithm. They wanted to figure out how these models could scale to the full neocortical scale and copy the functioning of minicolumns.

They also investigated a spiking neuron columnar model. They performed a literature review on a variety of neuronal software imitations including Hodgkin/Huxley (HH), Morris-Lecar, and Izhikevich. They built software to emulate aggregate neuronal function with a higher level of detail than some of other parts of this project. Most of the other things involve duplicating brain modularity at a level above that of the neuron.

Finally, the paper discusses a confabulation program that has been developed. Confabulation means that the program can make up completions to partially blank sentences. The researchers trained the sentence completion program by feeding it a lot of text. The intelligent text recognition would perform differently depending on what author it had read. For instance when it was trained using only material written by Shakespeare, the confabulated (made up) sentences resembled ones produced by that author. The "Original" sentence is shown below. That is followed by the "Starter" words "Go to" that are fed to the computer. The "Completion" is the made up sentence that is based on the two starter words.

Original: “Go to the forge with it then shape it I would not have things cool.”

Starter: “Go to”

Completion: “Go to me at your convenient leisure and you shall know how I speed and the conclusion shall be “

The scientists have combined a few of these discrete models into one cohesive program. They have developed a hybrid BSB/neuronal model and a hybrid BSB/confabulation model. They have also investigated about the potential of scaling up the hierarchical bayesian model and the fixed point attractor network models to an entire brain.

It appears that they are getting some interesting results thus far. The hybrid BSB/confabulation model was able to recognize character fonts using a 16 by 16 pixel array. They fed the program sentences with characters missing (see picture left). When 20% of the characters/words were missing, the program was 99% perfect in identifying the correct missing letters. This success rate was due to the program being trained on reading material.

I've been skeptical myself of being able to copy the functioning of the mind in software. However a lot of this work is more inspired by how the brain works, instead of imitating it exactly. I think this research on neuromorphic AI looks promising. If scientists can meld aspects of what the human brain does best with what computers do best, I think we will see a lot of interesting things come out of it.

Read more...

Neuromorphic Architectures Beyond Moore's Law

Neuromorphic engineering seeks to emulate how the brain functions via computer chips and/or software in order to better carry out specific functions. The brain really is an amazing organ than can do certain tasks much better than any computer. A lot of the research being done in this field is building models that are inspired by how the brain functions. The military, for instance, is keen on developing this technology for better AI learning/recognition of reading material among numerous other potentially beneficial tasks.

Moore's law itself doesn't appear to be ending soon, at least when it comes to increasing the amount of transistors. However even with this boost in computer chip hardware, certain software measures are just not improving at the same rate. Neuromorphic engineering has the potential to exploit how the brain functions in order to increase the speed at which specific tasks are performed (such as visual recognition). Here's a short abstract describing the potential (PDF).

Within this period traditional scaling of transistors in CMOS technology, will have reached its physical limits. However, advances in nanoscale structures such as carbon nanotubes and semiconducting wires have the potential to add new functionality by augmenting traditional processing in nano-CMOS technologies. We will move slowly but steadily towards an era where breakthroughs in the field, will not be driven only by research aimed at exploiting and managing the exponential rate of digital transistor density in CMOS technology (Moore’s Law). Research and development in the field will strive towards benefits from methodologies that allow advances in the structural complexity of micro-systems.

From an architectural viewpoint, we take inspiration from Nature. Biological information processing systems employ dynamic matter and learning at all levels in an amazing network of complex structures of different scales, from the nano to the micro and macro. Both biological brains and sensory organs operate at performance levels set by fundamental physical limits, under severe constraints of size, weight and energy resources and they are indeed engineering marvels of heterogeneous integration and structural complexity across different physical scales,

Another abstract discuss the potential (PDF) of using memristors to model neurons.

Memristive nanodevices may fill the role of an electronic analog of biological synapses: they are essentially analog memories that can be switched between extreme states in 20 nanoseconds or less, yet maintain their state for years when power is removed. They can also be manufactured at biological scale densities (more than 1010 devices per cm2) and integrated with conventional CMOS.

A workshop just recently took place to discuss these novel computing advancements.

Read more...