Benjamin Schrauwen

I love multi-touch technology and cannot wait until it becomes ubiquitous. But we haven't seen the last of it yet! This really cool technology demo of recent multi-touch research at Microsoft presented at Microsoft TechFest 2009, shows that stunning ideas are possible by stopping to think of a display as a 2D plane which passively sits in front of you on your desk. I'll put one of those on my wish-list.

The semantic web has been predicted for some time now to be just around the corner. It however still hasn't happened. Instead of waiting, the Drupal community has chosen to already start building the future. RDFa will be part of Drupal core, and an early demo demonstrated what the power of this could be.

It is surely true that incorporating RDFa into everybody's CMS is the way to get the semantic web kickstarted. But what will this mean? Sure, it will be easier to write mash-ups, easily extract structured content from websites, and work with wonderful RDF technology such as SPARQL, but it's my feeling that this will not motivate people to such a degree that they want to update the complete infrastructure of the web. I think it is only when there is real end-user usability improvement that the semantic web will really start to fly.

Happily enough, there are already some Firefox plugins available that start showing what might be possible in an microformat-ised or RDF-ised internet. These are the ones I tested:

Operator: Operator leverages microformats and other semantic data that are already available on many web pages to provide new ways to interact with web services. It is mainly focused on microformats, but it does support RDF. An additional toolbar appears which is context-sensitive, and provides actions dependent on the semantic data which is available on the current page. When for example contact or event information is available, actions appear that allows you to add it to your address book or calendar.

Zotero: Zotero is a free Firefox extension to help you collect, manage, and cite your research sources. It is able to automatically detect various meta-field in websites and helps to easily extract bibliographic information. This tool really does what it is supposed to do. Here you start feeling the power of what the semantic web could eventually be like. The trick behind the scenes is however that it doesn't really use semantic data yet, but has many custom content extractors for various often used websites. But it is due to this that you get a tool that at least makes the web feel like it is already a semantic place.

Tails export: Tails export is a Firefox extension extension for showing and exporting microformats. Similar to Operator, but it does not provide really sensible action, it just allows you to export the semantic data.

Sematic radar: Semantic radar is a semantic metadata detector for Firefox. It's focus is on RDF. The plugin appears to be more of a developer tool, as the provided operations on the semantic data are of little practical use.

Monty Python opened its official YouTube channel, yay! Apparantly, after opening the Monty Python channel, and freely giving away copyrighted content, their sales on Amazon increased by 23,000 percent!

I really adore Monty Python, and this sketch is really one of the best:

The president "almost-out" George W. Bush is quite a big spender. The national debt in the US has reached a record number during his presidency. Although Bill Clinton was since decades able to lower the national debt in the nineties, Mr. Bush happily spent a buck extra to compensate for it.

I.O.U.S.A. is a very well made documentary by Patrick Creadon which lays down the clear facts and history of the long mounting up debt crisis. Add to this the current financial crisis and the upcoming baby-boomers retiring wave, and you get a quite gloomy picture of the fiscal, and as a consequence political, future of the US.

This 30 minute preview of I.O.U.S.A. is well worth watching:

The Onion has exclusive coverage of the "the most intuitive product ever designed"! This revolutionary laptop without a keyboard is really the next big thing. So beautiful and simple, just perfect!

Apple Introduces Revolutionary New Laptop With No Keyboard

As most people know, there are 3 passive electronic devices: the resistor, the capacitor and the inductor. However, already in 1971, Leon O. Chua hypothesized, purely based on symmetry arguments, that there must be a fourth basic two-terminal component, and he called this device the memristor: it acts as a resistor which has memory. From a purely linear standpoint this device would not exist and would just act as a resistor, but it is thanks to Chua's focus on non-linear circuit analysis that the existence of this special device was predicted.

For many years this idea was thought of as a hoax and was well headed into the scientific dustbin of ignorance. It is only very recently that the people of HP Labs were able to manufacture a nano-scale memristor and published the results in Nature. These memristive properties turn out to be very abundant at the nano-scale level, and in retrospect have been reported in hundreds of paper over the last decades, but they were never understood and linked to the concept of a memristor. It was thanks to the years long effort of the HP research lab that they finally could link the anomalous behavior which was perceived at the nano-scale level with the ghostly memristor device.

The device HP actually build is a quite elegant and simple device: take two conductors (nano-wires in this case) and put an TiO2 isolator between them, but make the isolator so that at one end it is depleted of O2 ions. When a significantly high voltage is put over this device, the 02 ion holes gradually migrate towards the other side of the component, and because these ion holes are conducting, in this process the conductivity of the component increases. The real key part is that these ion holes are immobile when the driving voltage is removed, and the component thus remembers its resistance (apparently for a very long time). An inverse voltage again decreases the conductivity as this forces the ion holes to move back to one edge, and turn the device back into an isolator.

But why all this fuss over a new mystery device that only lives at the nano-scale? It is at the nano-scale and it has memory!! These devices are the perfect non-volatile memory components of the future. They could also be used to drastically improve the density of FPGAs. Another very important aspect is that they allow to redefine the way Boolean logic is implemented in computers. And what I find most interesting, there is a direct link between a memristor and a synapse in a neuron! There already have been circuits build which implement this. In a recent paper is was BTW also shown that memristive properties lead to learning in a very natural way and can explain how, for example, simple single cellular creatures can learn from their environment.

On November 21, 2008 a first full-blown symposium on memristive devices was organized at the University of California, Berkeley. The (quite long) four part videos of this symposium are available if you really want to get some in-depth information on these devices which have a great future lying ahead in the nano-scale electronics era!

At NIPS 2008, Lars Buesing received the NIPS student best paper award for our paper "On Computational Power and the Order-Chaos Phase Transition in Reservoir Computing". Woot!! Four out of the 1000 submissions got a student best paper award, which consist of a cash price and an invitation to send an extended version to the Journal of Machine Learning. Congrats Lars, you did a wonderful job on the paper.

NIPS was BTW quite stunning this year, an incredible improvement since last year! Finally NIPS again deserves the "Neural" in its name. For once not all paper were on Graphical Models, SVMs and Gibbs Sampling...

Companies such as tikitag (an internal start-up of Alcatel-Lucent) are desperately seeking for killer apps that will kick-start the RFID revolution. "The internet of things" have been predicted and evangelized about for quite some time, but the widespread use of RFID technology has still not happened. Many of the possible applications that are currently envisaged, are really not it (showing the website of "Toy Story" when you touch a RFID scanner with a Woody doll), and I have the feeling that these will not solve the current chicken and egg situation. But there is one thing which is gaining extreme momentum lately: people really want to live in a more sustainable way, but it is really hard for the average consumer to know what the ecological impact is of the products they are buying.

Alex Steffen presented in his CWF talk and in his WoldChanging.com book that one cornerstone of a sustainable world is metering. If people can easily meter what they use, they use less. A golden example is the fuel efficiency meter in the Toyota Prius: adding a simple gauge even increases the efficiency of the Prius because people try to be as efficient as possible.

This could also happen for everyday products. I propose to add RFIDs to everything. Not just on expensive products, but really on everything you buy. The main purpose of these small product-following-memories is to track the ecological history of the product. In the chips memory would be stored how much carbon dioxide was dumped to produce the product, how much water was used, the amount of toxins that were excreted, ... This bookkeeping should not only be kept for end-user products, but also for every conceivable base product. By setting up this eco-accounting system from the (almost literally) ground up, producers, transporters, and resellers of products can easily calculate the effective end-product footprint.

You might think that such a ground-up tracking of products is not feasible, but after years of scandals in the production of meat and dairy products in Belgium (mad cow disease, dioxins, bird flue, ...), there now has been enforced a complete traceability of these products. And this system actually works, but is still based on old technology such as burn-marks, stamps, and piles of paperwork ...

If we want to really make a change, we need not (only) tax the CO2 footprint of the industry, but put the eco-taxes directly at the expense of the end-user. Let the end-user pay the bill for CO2 heavy products. By doing so, we make CO2 a directly marketable product. Economical free-market dynamics will do the rest. One key technology which is needed to achieve this is ... micro CO2 accounting, which is provided by the eco-RFID revolution.

Oh and yes, by the way, this whole "internet of things", you just get it for free.

This November 21 2008 Google tech talk of Karl Deisseroth on optogenetics is quite stunning. The idea of optogenetics is to genetically alter specific types of neurons of alive and behaving subjects (currently rats) with genes coding for opto-sensitive ion-pumps (ion-pumps are small proteins or "bio-machines" which reside in the membrane of the cell and which pump certain ions for the cell's surrounding to the cell and dump other ions out of the cell). The genes that are added to these specific neuron-types, are taken from simple single cell algae and microbial life where they code for optically driven ion pumps used by these simple lifeforms to directly extract their life's energy from photons. These optically driven ion pumps are then artificially placed in specific neuron cells, which allows to optically excite or inhibit very specific brain regions by, thanks to the bacteria's ion pumps, effectively injecting an electrical current in these cells by simply shining light on them. This incredible techniques allows for an unprecedented precision in the external control of neurons in the brain. As an example they show a modified rodent which they can externally steer, by simple pulses of light (the controlled rat really desires to walk in leftward circles). But this toy example really pales when the true far-reaching consequences of this technique are understood: it introduces a whole new ballgame in how we can start studying the exact operation of the brain up to the tiniest details, and it can provide cures for diseases such as depression, epilepsy, Parkinson's and Alzheimer's disease. I started studying Deisseroth's papers today.

I've always been extremely fond of everything synthesizer related: analog synths, software synths, modular synths, ... I experimented with and build all of these types of synthesizers in my younger years. At Ghent University I even was once adviser of a master thesis where the students have built a modular digital synthesizer on FPGA.

But all of this fades into oblivion if you see the Reactable, research from the Music Technology Group - Pompeu Fabra University - Barcelona. It is a mix between Jeff Han's multi-touch display, an intuitive and tangible modular synthesizer, and a collaborative music instrument. This is something I would definitely want to own (and build, which is half of the fun :). The software powering the Reactable is open-source by the way! Very cool indeed...