Spend enough time in the Arctic, and the featureless vista of white snow can eventually make you go blind. The reason is that if you stare into blank nothingness long enough, your brain shuts down your visual processing. Your eyes simply cease to carry the unchanging input. You may even start to hallucinate. It’s called the Ganzfeld effect, and it highlights how changing and fluctuating external input — or input noise — plays a role in the functioning of the brain.
So what is this noise all about? Well, for starters, there’s noise in your brain. Just like a TV set that’s not tuned into a channel, your brain produces static when it’s got nothing to do. This intrinsic noise is actually your neurons firing at random. And when external noise (that is, input fluctuations) is introduced, there is an interaction between the two kinds of noise. Arvind Kumar, a researcher in computational biology at KTH Royal Institute of Technology, says that to some extent, this interaction holds the key to how our brains function.
“You have noise in the brain when you aren’t doing anything; and when you start doing something, this noise is reduced,” he explains. “But it is not well understood how this noise is reduced even though overall activity of the brain increases.”
But let’s return for a moment to our Arctic explorer. The neurons in his or her brain become elevated as they seek in vain for some correlation in the visual input. It’s a fruitless effort because, unlike in a forest or on a city street, there are no correlations in the visual stimulus you get while staring into a blank void. A flat wall has no structure the neurons can get meaningful input from. Look at it, and the result is random electrical fluctuations in your head, as the brain gives up on processing the input.
“We want meaningful inputs,” Kumar says. “And this will have fluctuations around an average value. But these fluctuations are structures. For instance, when you look at a landscape, different pixels have this structure in terms of their relationship to each other.”
Everywhere in our world, there is noise — or more accurately, correlated fluctuation —inside and outside of our brains. And when these two noises interact, a curious thing happens. Instead of elevating our intrinsic noise, noisy input appears to cancel it out.
The mystery is, why?
One theory is that there are some mechanisms within the brain that quiet down the noise internally. One such mechanism could lie in the brain’s wiring. “That explanation is compelling”, says Kumar, “but it’s incomplete. If the mechanism is hardwired, quenching of the noise would be independent of the stimulus,” he says. “The very cause of the change in the brain activity, that is, the stimulus, has to play a role in determining the response.”
In a new paper, published in the Journal of Neuroscience, Kumar argues that the fluctuations in the input (or the external noise) itself is what can cancel out the intrinsic noise of the brain.
“The input itself has properties that cancel the noise,” he says.
It goes something like this. If you look out at a river and there are waves, and you throw a small stone in, you wouldn’t see its effect on the wave structure. But if you throw a big enough stone, you see the change clearly. That is, the input must have features that are bigger than the noise in the system.
More specifically, Kumar and his colleagues point out, some input features mainly affect individual neurons, while others affect groups of neurons. Input features that affect individual neuron responses determine the variability and noise in the response. Input features that affect populations of neurons define the magnitude of the response.
Kumar says that with this study, there are two plausible mechanisms that could reduce the noise and variability in the brain – the external input, and the local connectivity in the brain. Kumar says that the ultimate explanation may be a combination of a noise dampening mechanism inside the brain coupled with the affect of external noise.
“I think this is what makes biological systems complex and interesting, that when you have multiple explanations, all are correct to some extent. The challenge is to find out the domain of their applicability.
“Fortunately, in this case the two explanations make different and testable predictions,” he says. “So, the ball is back in the court of the experimentalists to test these theoretical predictions.”
From the moment the Jetsons’ robot maid, Rosie, smashed a pineapple upside down cake over Mr. Spacely’s head, I’ve looked forward to the day when humanoid housekeepers would inhabit our homes. But interacting with humans is harder work than it seems. So, we can forgive science for taking a bit of time with developing artificial intelligence.
One of the most complicated aspects of artificial intelligence is getting machines to understand what we want. Language would seem to be the best way to direct a robot, but the truth is we’re not even sure how humans understand language.
Machines can learn language to some extent, but it is harder to program the context and experience we rely on to not only understand language, but the things we encounter as well.
A program can enable a computer to see a picture of a bottle and then say what the picture is. But Christian Smith, an assistant professor in Computer Science at KTH Royal Institute of Technology, says that it’s a “completely different thing to see a picture and understand what you can do with it.
“I can program a robot to pick up a bottle of water and pour water into a glass. But it might not be able to use that knowledge to pour gasoline from a container into a car. It’s basically the same kind of action, but we cannot yet get robots to generalize those kinds of things.”
The fact is, it won’t be able to pour from just any bottle either. A robot can only be programmed to pour from a specific bottle, which is fine if you want all the drinks in your home to be decanted from the same, standard bottle (just be careful your robot’s not serving bourbon in place of your morning orange juice).
And while there are language processing dialog systems for computers, these may not necessarily work well in an interaction between a human being and a robot.
How hard can it be? Smith points out the challenge for a machine to understand context. Take for example the determiners, “this” and “that”. You’re sitting in the kitchen and there are two glasses on the table. “I have this glass, and over there we have that glass,” Smith says. “But geometrically speaking, where is the boundary line where a glass stops being this glass and starts becoming that glass? If I tell you to pick up that glass, which glass is that?
Now, imagine you’re sitting at a table with another person. There are two glasses in front of you, at separate distances, and you’re asked to pass “that glass”.
The one that’s closest to you will be this glass, and the one farther away from you will be that glass, even though they’re both that glass to the person on the other side of the table.
“The glass that is that glass in one context becomes this in another context. This and that will point to completely different glasses just because of our spatial context.”
Sounds like the premise for an Abbot and Costello routine, right?
How does that person know which glass you are referring to? Usually we reach agreement with the person who’s sitting at that table with us. Eye movements, gestures and body language help two people agree on which glass they’re talking about; though I’ve been in situations with humans who have normally functioning brains (at least I suppose they did) and still got it all mixed up. So I can’t help but imagine how difficult this could be for a robot.
However, it is becoming increasingly clear to me that we are a long way from having humanoid robot bartenders.
David Callahan
Watch a full discussion of the challenges and issues around artificial intelligence, on Crosstalks TV, featuring Smith along with: Jürgen Schmidhuber, Professor in Artificial Intelligence, Scientific director at the Swiss AI Lab IDSIA; Theo Kanter, Professor of Computer Science at the Department of Computer and Systems Sciences Stockholm University; and Kristina Nilsson Björkenstam, PhD, Computational Linguistics Stockholm University.
Crosstalks is an academic web talk show where recognized researchers from two of Sweden’s top universities, KTH Royal Institute of Technology and Stockholm University, discuss global topics live with viewers worldwide. Crosstalks is an international academic forum where the brightest minds share knowledge and insights on the basis of leading research.
Andreas Rehn explains the idea behind the Zone 164 reading pod. (photo: Morgan Waltersson)
A team of graduate students from KTH Royal Institute of Technology went to the Stockholm suburb of Kista last year to find out what 13- to 15-year-old boys think about books. Of the 38 boys they interviewed, only five said they had ever read a full book on their own—aside from the autobiography of Sweden’s football superstar, Zlatan Ibrahimović.
Considering all the competition for their attention, all the instant gratification readily available via movies, video games, the web, it’s a wonder kids immerse themselves in anything that requires intellectual effort. But a lot of kids do. And if they read books, there’s a good chance it’s happening because these kids are encouraged to do so. Not just encouraged in the sense that the parents started reading to them when they were infants, but encouraged in the way that their parents provide an environment where kids can concentrate.
Which sounds perfectly reasonable for a parent who is not raising four kids in a 60-square-metre two-bedroom.
That’s the story in a lot of homes in Kista, a low-income suburb that is home to many families who fled to Sweden from war-torn, developing countries. A lot of the parents here have big families to raise, and they do so in small apartments where no one gets much privacy.
Lacking a place they can call their own, the Kista teenagers tend to drift out in search of environments where they won’t be pestered by parents and siblings, says Andreas Rehn, a master’s candidate in media technology. But the places where boys in particular wind up in are not exactly peaceful and solitary.
The City of Stockholm wants more teenage boys (among others) to spend time at the Kista Library; but the dual — and sometimes conflicting — desires for privacy and immersive stimulus rarely merge into a desire to spend time there. So the city asked OpenLab, a interdisciplinary design thinking laboratory housed at KTH, to come up with a solution.
The reading pod’s dimensions.
What Rehn and his team came up with combines stimulation and the sense of one’s own private space. It’s called the Zone 164 (the postal code for Kista) reading pod—a partially enclosed, cocoon-like chair that is fitted with audio and light outputs that can be engineered to deliver subtle enhancements to the actual text that the user is reading.
Rehn and colleague Viktor Wennström built the prototype for the pod together, using the development platform, Arduino. Also in the group were Anna-Belle Ericsson, Sarah Eriksson, Johanna Thuresson and Michaela Woltter.
“We want to lower the threshold for reading,” he says, as I settle inside the prototype sitting in a corner of an OpenLab workroom. “We wondered, is it possible to create an overlay over traditional reading that stimulates the senses that they are stimulating now with different kinds of media? What can we take from movies and music?”
Light and sound, to start with. Colored lighting and ambient background soundtracks are what are currently installed in the prototype. I’ve put it on the night setting. I am surrounded by a deep violet glow and the sound of crickets chirping and frogs croaking in some forest. And it is quite cozy and muffled—without being claustrophobic because, the rest of the world is clearly visible on one side. And the nice thing is, I don’t hear much of the rest of the world.
“It gives you some distance from what is going on around you,” Rehn says. “It blocks out the peripheral vision. That’s helps the user to concentrate.”
Rehn tells me it could be more interactive. There is existing technology for embedding book cover chips with custom lighting and sound cues that can be activated by eye-tracking sensors.
Yet the idea isn’t to overwhelm the reader. “This is passive feedback,” Rehn says. “The light is bouncing off the paper and the page becomes tinted.
“We don’t want to distract the reader; we want to heighten the experience and trigger the senses.”
And then he suggests something that immediate triggers mine. “These could also be used in open seating work environments.”
At which point, I begin to image little clusters of these cubbyholes scattered around my own office; and I start to fully understand how truly awesome this actually is.
The second, successful launch of ESA’s Cluster mission in 2000 was the beginning of something that outlasted expectations. (Photo: ESA/Starsem)
Like the Duracell Bunny, the four Cluster spacecraft that ESA sent into space 15 years ago have continued working long after anyone expected them to. And in its extended lifespan, the mission has paid dividends. Over the last decade, transmissions of data from Cluster enabled a scientist at KTH Royal Institute of Technology to make a major contribution toward the understanding of Earth’s aurora.
By any measure, Cluster has been successful. But it’s an especially rewarding outcome for a mission that was set back by a rocket crash on its first launch attempt. The original four Cluster satellites never made it into orbit. The four replacement satellites, which are still in use today, were launched aboard Soyuz-Fregat rockets in 2000.
An artist’s depiction of two of the Cluster satellites in orbit. (ESA/J-L.Atteleyn)
Space missions are typically guaranteed to work for a few years, during which time all the primary scientific objectives are supposed to be met. Per-Arne Lindqvist, a space and plasma researcher at KTH, says that unmanned craft often continue working longer. Yet the fact that Cluster consists of four vehicles and has been in operation a full 13 years beyond its original expiration date has been somewhat of a pleasant surprise.
Even stranger is that by surviving so long, the four spacecraft have actually been able to get a better view of Earth’s aurora, which made possible the recent breakthrough in the modeling of the inner workings of the so-called “black auroras”.
KTH contributed to Cluster’s detectors for measuring electric fields and waves (EFW); and the university hosts the Scandinavian data centre for analysis and distribution of EFW data. (More than 2,000 publications of Cluster results have been released, of which 250 come from the EFW team). Cluster’s mission was to observe Earth’s magnetosphere—the invisible bubble of highly-charged electrical and magnetic fields surrounding the planet. But, as a result of its long lifespan, it wound up doing a little bonus work:
“The main thing about Cluster and the aurora is that the original orbit was such that it was not best suited for studies on the magnetic field lines where the auroral acceleration takes place,” Lindqvist says. “Over the years, due to natural disturbances, the orbit has evolved in both shape, size and orientation, so coverage of the auroral field lines became much better after the original mission duration. And of course, data over several years enable statistical studies.”
Beginning about two years after the original Cluster mission expiration date, Tomas Karlsson, a space and plasma physics researcher at KTH Royal Institute of Technology, started to notice something strange in the data being sent down from the project’s four satellites.
“I noticed a weird combination of electrical and magnetic field measurements that were different from normal, and I wanted to understand the physics behind the data,” Karlsson says. “On each occasion, the Cluster spacecraft were flying over the night-time auroral region.
“These events are very rare, and if I had only seen it once, I might not have attached any importance to it,” he says. “But when I found a number if similar events, over the years of data, this made me think that they were relevant to look closer at.”
And then there is all of the follow-up study of the results that continued at KTH and other universities around the world after the first two years of the mission.
ESA has extended the project in two-year increments ever since 2002, and while there is a probability that the project could extend beyond 2016, that probability is decreasing all the time. “The main problems on the spacecraft side are power to operate everything, and fuel to keep the spacecraft orbit constellation and attitude,” Lindqvist says.
And the instruments are not going to run forever. Karlsson says they are degrading at different rates on each of the four spacecraft. “Some of our electric field probes have failed at different times on different spacecraft,” he says. “We can still do good measurements even if we lose two out of four probes on a satellite, but with poorer time resolution.
“If one spacecraft would fail completely, I guess there would be serious discussion about a continuation. But one must also consider the added value to other missions, such as the upcoming MMS mission (which is in orbit and being commissioned now), to have the Cluster spacecraft to correlate measurements with.”
Lindqvist adds that the only thing keeping the power going in the Cluster spacecraft now is the Sun. “All batteries have been dead for a few years, so the spacecraft operate only in sunlight. But, the solar panels have actually degraded much less than anticipated. The fuel is running out but there is still a small amount left.
Physics researcher Egor Babaev speaks from experience when he tells his students that rejection can sometimes be taken as a form of compliment.
His most influential paper, published in 2003 on the subject of superconductivity, was initially turned down by a number of journals on the grounds that its conclusions couldn’t be possible in principle. It took Babaev two years to get it published, but the effort was worth it. “It is one of my most cited papers, but at the same time it was the hardest to publish,” says Babaev, an assistant professor at the Department for Theoretical Physics at KTH Royal Institute of Technology. It also led to a prestigious award from the Royal Swedish Academy of Sciences.
As a postdoctoral researcher at Uppsala and Cornell universities, Babaev had proposed a type of superconductor that didn’t fall inside of the traditional classification of superconductors. Neither type I or type II, the type 1.5 superconductor he predicted has not only garnered numerous citations, but research funding from the U.S. National Science Foundation.
Among those citing the were prominent researchers from Princeton and Stanford. “I got many invitations for seminars,” he says. “I tell my students about it as an example: if you try to do something innovative, do not expect referees to immediately and uniformly accept it.
“When we get occasionally get a similar reaction from a referee, that can be taken as a compliment,” he says. “In physics one should try not to take for granted, but challenge, even the most basic notions, no matter how well-established they may seem to be.”
David Callahan
Watch Egor Babaev give a presentation on type 1.5 superconductivity
Author
