This decade, particularly the second part, will be characterised by cross fertilisation deriving from the application to one domain of science of discovery and paradigms coming from other domains.

The last century was a century of specialisation and layering; typically visible in the ICT world where you see great specialisation and the progress of one discipline fuelled by the progress of another because it is layered on it: the progress in information technology is largely the result of progress in electronics, more processing and storage capacity have fuelled progress in ICT, but the electronic specialists have no say in information technology and vice-versa).

Main brain areas involved in the honeybee flight

The next decade will see paradigms developed in one discipline to change the approach in others.

An example can be seen in this news of the Green Brain project, driven by researchers at the Sheffield University. Their goal is to “copy” a honeybee brain to a tiny robot for having it flying with the same agility of a bee (and they are pretty good at it!).

Clearly there is the mechanical challenge of creating a tiny robot with the appropriate moving parts (and we are still far from dominating this technology, researchers have managed to create a robot that is cumbersome if compared to a bee, as shown in the photo below.

However, once you have a robot you need to provide it with the controls that would allow it to fly and negotiate obstacles, find the good path, take into account any small variation in air flow (they are crucial at this size…) and applying the same aerodynamics of a plane to a tiny flying robots does not work.

Hence the attempt to copy a honeybee brain….

Actually copying it all would be too big a challenge for today’s science and technology and it may be beyond the goal. What scientist are trying to do is to understand how the honeybee interprets the signals coming from its senses (vision, smell, accelerometers, position, airflow) and convert them into action for the flight.

In the drawing you can see the various areas of the honeybee brain in relation to this task.

A robot bee. As you can see it is still bigger than a bee and the wing span is considerably larger

In the near future we are likely to see more and more research to understand how Nature has solved tricky problems by evolving sophisticated control centres (brain) and to replicate them to solve our problems.

At the same time miniaturisation technique, smart materials and nanotech will likely provide the tools to build micro robots. As noted, we cannot just scale down our science for big planes to these dimension, a brand new approach is needed and copying Nature may be the best way to do it.

As the residential user has access to greater bandwidth, new services that previously were not possible, are now closer to reality. Among such services, it is possible to highlight applications that make use of holography resources. The applications areas are potentially many and among them I feel that the one of distance learning can be very interesting. The end user, in this context a student, could interact with holographic objects such as molecules and plants in order to learn by using virtually realistic projections of these objects. This type of visualization would provide further access to scenarios that are not possible or feasible in a conventional classroom, for example, to know details about endangered animals, or plants by a projection of itself with a previously unseen level of detail and this is what may trigger a more profound understanding and speed up the learning process.
Recent studies move the fiction into reality: the technology required to provide the recording of holographic projection and dynamic 3D modeling already exist although the visualization of holographic object is still a big issue in terms of cost and object dimension.
Another area that could take advantage of holographic visualization would be the gaming space. Solve an IQ Teaser, for example, a Rubik’s Cube, interacting directly with holograms that are projected in front of you would be a better way of engaging your brain.

HP Labs has announced (New York Times Nov. 18, 2009) a project called “Central Nervous System for the Earth” (CeNSE): it is a R&D program to build a planet-wide sensing network, using millions and millions of tiny, cheap and sensitive detectors (e.g. accelerometers that detect motion and vibrations, ones for light, temperature, barometric pressure, airflow and humidity) [1]. CeNSE is another effort in building a pervasive infrastructure gathering and processing huge amounts of data.

We see a general consensus that data around the world will continue growing and everything will become more instrumented and interconnected (Internet of Things). Most probably, this trend will bring also the need of innovative computational and networking paradigms – i.e. enhanced with cognitive capabilities capable of spotting patterns in huge amount of data, to analyze, route and integrate information real-time and to deal with stochastic behaviors typical of complex, real environments. A question arises: are current computing paradigms ready to face these challenges?

In the direction of inventing brain-inspired computational paradigms (facing challenges of a pervasive Internet of Things), a recent press released [2] announced that scientists, at IBM Research, in collaboration with the Lawrence Berkeley National Lab, have performed for the first time a near real-time cortical simulation of the brain (with a scale of 1 billion spiking neurons and 10 trillion individual learning synapses).  Specifically, in collaboration with Stanford University, IBM has developed an algorithm (Blue Matter) [3] exploiting the Blue Gene supercomputing architecture in order to map the connections between all cortical and sub-cortical locations within the human brain. Final goal is building brain-like computing systems based on a compact, low-power chip using nanotechnology and advances in phase change memory and magnetic tunnel junctions.

On the other hand, as Roberto pointed out in November 21st post, new results seem indicating that some elementary brain-like functions, e.g. counting, may be achieved with only a hundred neurons, that advanced thinking can be done with a greater but still limited number of neurons and that even consciousness might be generated (in principle) with smaller neuronal networks than billions spiking neurons! Nature is amazing.

Reading that I am more and more convinced that a clear understanding of how the brain works requires considering higher level models than the ones at the neuronal level. This is what Nobel Prize Prof. Brian D. Josephson argues in [4]. It looks reasonable to think that subtleties of brain such as consciousness have as basis higher level forms of description than ones at the neuronal level, whose role might be limited to justifying the higher levels. For example, Prof. J. Cowan (University of Chicago) discovered that neural activity could be effectively described as an analog of a chemical reaction-diffusion process (see my last post), and that at rest, it is statistically similar to Brownian motion; such description ignore, in turn, first-principles derivation from very basic physical laws. Nature has jumped ahead of us: once we become aware of these generic biological models we can probably make use of them to progress our understanding of the brain.

This approach leaves main questions open: What are these generic models? How do hierarchies of models emerge (and concatenate) in the brain? How are higher level models justified by the neuronal level? And beyond that, at the very basic physical laws?

References

1. http://www.kurzweilai.net/email/newsRedirect.html?newsID=11423&m=54146
2. http://www-03.ibm.com/press/us/en/pressrelease/28842.wss#release
3. http://www.modha.org/C2S2/2009/11182009/content/MICCAI_BlueMatter.pdf
4. http://www.tcm.phy.cam.ac.uk/~bdj10/mm/articles/cognition.html

It has been believed for the last 50 years that our “human” brain can be so smart because it has over 85 billion neurons intricately connected. It is this extremely complex web that gives rise to intelligence and consciousness.

Newer studies carried out by researchers at the Queen Mary University in London are now pointing out that tiny brains, such as the ones of insects may be as smart as much bigger brains.

Testing the IQ of a Honeybee

Queen Mary’s researchers have shown that honeybees can count, categorize similar objects like dogs or human faces, understand the concept of “same” and “different” and perceive what is symmetric and what is asymmetrical. Now that is not like reading Goethe but it is a level of smartness that was not believed possible in such small brains. A honeybee brain weighs only 1 milligram (compare it with our 1.3-1.4 kg or a whale, 9 kg) and has less than a million neurons.

http://www.sciencedaily.com/releases/2009/11/091117124009.htm

Researchers have found that the size increase of a brain let it work with more precision, higher sensitivity and so on. Clearly it helps in being smarter. However, bigger brains are also bigger because they need to operate bigger bodies, they have more circuits but they are repeating one another. They can control more pieces but not necessarily they do that in a smarter way. According to researchers a bigger brain compares to a bigger hard drive but does not necessarily stands for a better processor.

The new results seem to indicate that counting may be achieved with about a hundred neurons, advanced thinking can be done with a greater but still limited number of neurons and even consciousness can be generated with small neural circuits (as little as a few thousands according to their research): are honeybee conscious? We do not know, yet. But this is quite a change from the assured NO we used to respond to that question.

The results from this study, if confirmed, is important because we almost have the technology to create, in silicon, a honeybee sized brain. The target is to reach that kind of complexity by 2011.

If the hardware is sufficient, than the question is about the software to exploit it. If we find out the way honeybees can distinguish faces and categorize objects in the environment we could replicate that to distinguish images much better than today. iPhoto 9 is attempting (and basically failing) to do smart face recognition. May be iPhoto 10, embedding a bit of honeybee savvy may succeed.

Researchers at the Arizona State University have found a way to create a diode (the bit equivalent in electronics) using a single molecule. The drive towards smaller and smaller elemental component has been on for the last 40 years. Now miniaturization has reached the nanoscale (40nm is now an industrial process with 20nm on sight in the next decade and possible 4nm by the end of the decade). A single molecule is below the nm size but one has to remember that to get the number of molecules needed at a, say, 20nm scale one has to consider that they are arranged in a volume, non on a line. Hence, to create a component based on a 20nm scale with a 0.2nm molecules you need 100³ molecules, that is 1 million molecules. Note that this calculation needs to be taken with a grain of salt since many of those molecules also serve the purpose to create a substrate, still we are talking, even at that incredibly small size of a huge number of molecules.

It comes as a staggering advance, then, the results of these Arizona researchers, led by N.J.Tao, that have shown how to create a diode using just a single molecule. The feat is accomplished by using asymmetrical molecules that respond differently depending on the interaction applied.

Asymmetrical molecule used as a diode

The technique developed by Tao’s group relies on a property known as AC modulation. “Basically, we apply a little periodically varying mechanical perturbation to the molecule. If there’s a molecule bridged across two electrodes, it responds in one way. If there’s no molecule, we can tell.”

It is interesting to note that this group of researchers operates in the Biodesign Institute, that is the department looking at how Nature works.

The application of these discoveries are still far away (10 years?) but it is nice to know that we have still a way to go ahead of us.

http://www.sciencedaily.com/releases/2009/10/091013110042.htm