Building Intelligence

There seems to be a growing number of scientists addressing the structure and functioning of the brain. I've been reading On Intelligence by Jeff Hawkins (the entrepreneur who started both Palm and Handspring), which builds a very convincing case for the brain building invariant models which it then uses in a pattern/prediction sequence.

I've also been reading Steve Grand's book "Growing Up with Lucy" - which claims to be 20 steps to building your own android, but is really a discussion about the practical implementation of AI and robotics. And, in turn, this led me to Cordelia Fine's excellent book "A Mind of its Own" - which has to be one of the more unnerving books I've read recently because it presents a very plausible case for the hypothesis that your brain lies to you most of the time - just to keep you cheerful...

I put together some brief notes in MindMeister - which is an OnLine brain storming tool that we've started using at Radley.

Follow me on Twitter: @IanYorston

Decision-making is still a work in progress for teenagers

It may not be wise to allow teenagers too much say in their educational futures...

Studies of brain development in teenagers may finally give parents the scientific authority to say "No you're not!" in answer to the common adolescent complaint, "But I'm old enough to make my own decisions!" That authority comes from brain imaging studies that reveal some surprising features of the adolescent brain.

Deborah Yurgelun-Todd and colleagues at the McLean Hospital Brain Imaging Center in Boston, Massachusetts have used functional magnetic resonance imaging to compare the activity of teenage brains to those of adults.

The researchers found that when processing emotions, adults have greater activity in their frontal lobes than do teenagers. Adults also have lower activity in their amygdala than teenagers. In fact, as teenagers age into adulthood, the overall focus of brain activity seems to shift from the amygdala to the frontal lobes.

The frontal lobes of the brain have been implicated in behavioral inhibition, the ability to control emotions and impulses. The frontal lobes are also thought to be the place where decisions about right and wrong, as well as cause-effect relationships are processed. In contrast, the amygdala is part of the limbic system of the brain and is involved in instinctive "gut" reactions, including "fight or flight" responses. Lower activity in the frontal lobe could lead to poor control over behavior and emotions, while an overactive amygdala may be associated with high levels of emotional arousal and reactionary decision-making.

The results from the McLean study suggest that while adults can to use rational decision making processes when facing emotional decisions, adolescent brains are simply not yet equipped to think through things in the same way.

Link: BrainConnection.com.

Follow me on Twitter: @IanYorston

Because.

A rather wonderful Introduction to the concept of Axioms. Well worth a read. And all credit to Robert G. Brown, of Duke University Physics Department, who wrote it.

Humankind has, from the earliest glimmerings of sentience on, endeavored to answer certain questions. What time is dinner? What's for dinner? Who caught dinner?

Once these questions were being answered satisfactorily and moderately regularly for at least a fraction of the proto-human population, we can only imagine that in that warm lull that follows a full belly around a fire beneath the stars, questioning thoughts turned to less important, but nevertheless intriguing, issues. I wonder what those little bitty lights are up there? I wonder how the fire comes out of chunks of cold rock or rubbing sticks? I wonder if I'll get dinner tomorrow, or turn out to be dinner for something else?

Eventually, humankind developed rudimentary societies, and for the first time at least the ruling class of those societies experienced the luxury of not having enough to do. Others in those societies didn't have that luxury but wanted to. Together they conspired to extend these questions still further, so that they could spend their overabundance of leisure time working hard trying to answer questions that had little to do, at first glance, with dinner.

[...]

We feel like all questions should have answers, but when we look at them closely we find that certain questions, of certain kinds, don't. Here is one:

- Why should all questions (including this one) have answers?

This is the kind of obnoxious question that parents eventually learn to answer (like all the rest of the related questions and question chains) with the simple word "Because.'' - which is, as any small child rapidly learns, code for "this question cannot be answered''.

"Because.'' as an answer to a "big'' question often means that it isn't a question at all. At least, not one that we can answer in the same way that we can give somebody the time upon request.

Such a "question'' calls into question - sorry - the foundation of what we know, what it means to ask questions at all, and how we should cognitively interpret the answers (such as they are, if they exist).

All of which leads us, by a roundabout way, to consider the issue of axioms.

Follow me on Twitter: @IanYorston

Home-Made Brain Imaging

Ah yes. Control of my hardware just by thinking about it... Exciting but scary.

The OpenEEG project is encouraging independent research into electroencephalograms (EEG), which provide visual representations of the electrical activity in the brain, by providing instructions on how to get EEGs on your PC.

The technology has been around for a while, but is not publically available because of standard IEC601, which requires medical devices to undergo specific tests before being made available, tests that are really expensive.

Luckily, you can now build your own! All you need is a signal capture card, two electrodes and some software (which is available free on OpenEEG's site along with comprehensive guides). Cheap capture cards are available through Olimex, electrodes are easy to find, and different software is readily available

Link: dreadful dreams.

Follow me on Twitter: @IanYorston

Unlocking How Flocks Stop, Turn, and Swirl in Unison

It's one of the wonders of the natural world—to see a flock of starlings pulse, wheel, and ripple as one across an evening sky. Just how do they perform these displays with such precision?

The question thas puzzled scientists for centuries, since many group-living animals have this talent for moving together in a seemingly spontaneous yet highly coordinated way. Anchovies, for instance, are as synchronized as starlings in their underwater ballets, especially when animated by the presence of predators.

Scientists had thought such graceful mass movements could only be achieved through complex signals. In the 1930s some even suggested these animals must be able to instantly transfer thoughts to one another. Or, in the case of birds, perhaps they follow a leader via electromagnetic signals.

Now a new study suggests there's a much simpler and more democratic navigation system that allows flocks, shoals, and herds to travel in unison.

For the study, published in the science journal Nature, researchers used computer models to show that large numbers of animals can move together relatively easily, even when few individuals know what's going on and there is no obvious leader.

[...]

Using computer simulations, the study team found that group coordination arises from two factors: the need for a group to stick together and the desire of some individuals to make their own minds up about where to go.

The researchers first programmed their virtual animals with a basic urge to stick as close together as possible without actually colliding. This instinct caused animals to form close-knit, evenly spaced groups, as seen in real mammal herds and fish schools.

Some animals were then programmed to have a preferred direction in mind, as if on a migration route or heading toward a particular feeding area. These animals' desire to reach their goal was then balanced with an urge to stay with the main group.

During these simulations, the researchers found that it needed only a few individuals to set off together in a certain direction for hundreds of others to follow.

Link: nationalgeographic.com.

Follow me on Twitter: @IanYorston

Brain Computes in Trinary not Binary

How come Nova always spots this smart stuff?

Interesting, interesting. The key point here is the idea that ignoring information may be a vital part of neuroscience... and that may have important implications for Artificial Intelligence.

MIT neuroscientist, Guosong Liu, has found that human neurons compute in trinary, using signals that are the equivalents of -1, 0 and 1. By contrast, all computers compute in binary, using just 0 and 1. Because the units of trinary computation can in some cases be additive (e.g. 1 1=2) or can "cancel out" (e.g. -1 1 = 0), the human brain is able to ignore information during computation, says Liu, something which present computers cannot do. Liu believes the ability for trinary computations to cancel out in some cases will enable next-generation computers to ignore information, and this will fundamentally change computing as we know it.

Link: Minding the Planet: New Finding: Brain Computes in Trinary not Binary.

Follow me on Twitter: @IanYorston

How many variables can humans process?

Boing Boing report that a new study shows that humans can usually track just four mental variables when trying to solve a problem.

In the journal Psychological Science, cognitive scientists from the University of Queensland and Griffith University report on a study where they tested these limits of processing capacity.

It's tough to measure because people commonly break down complex problems into manageable chunks. For example, a baker doesn't have to think: "break egg one into bowl, break egg two into bowl, etc." Instead, he'll track it as one chunk: "add all the eggs." To measure their test subjects, the researchers devised problems involving statistical interactions between fictitious variables. The details of the test are vague, but apparently the problems couldn't immediately be broken into "bite-size chunks."

The researchers found that, as the problems got more complex, participants performed less well and were less confident. They were significantly less able to accurately solve the problems involving four-way interactions than the ones involving three-way interactions, and they were (not surprisingly) less confident of their solutions.

And five-way interactions? Forget it. Their performance was no better than chance.

After the four- and five-way interactions, participants said things like, "I kept losing information," and "I just lost track."

Link: Boing Boing: How many variables can humans process?.

Follow me on Twitter: @IanYorston

Revenge of the Right Brain

The future no longer belongs to people who can reason with computer-like logic, speed, and precision. It belongs to a different kind of person with a different kind of mind. Today - amid the uncertainties of an economy that has gone from boom to bust to blah - there's a metaphor that explains what's going on. And it's right inside our heads.

Scientists have long known that a neurological Mason-Dixon line cleaves our brains into two regions - the left and right hemispheres. But in the last 10 years, thanks in part to advances in functional magnetic resonance imaging, researchers have begun to identify more precisely how the two sides divide responsibilities. The left hemisphere handles sequence, literalness, and analysis. The right hemisphere, meanwhile, takes care of context, emotional expression, and synthesis. Of course, the human brain, with its 100 billion cells forging 1 quadrillion connections, is breathtakingly complex. The two hemispheres work in concert, and we enlist both sides for nearly everything we do. But the structure of our brains can help explain the contours of our times.

Until recently, the abilities that led to success in school, work, and business were characteristic of the left hemisphere. They were the sorts of linear, logical, analytical talents measured by SATs and deployed by CPAs. Today, those capabilities are still necessary. But they're no longer sufficient. In a world upended by outsourcing, deluged with data, and choked with choices, the abilities that matter most are now closer in spirit to the specialties of the right hemisphere - artistry, empathy, seeing the big picture, and pursuing the transcendent.

Beneath the nervous clatter of our half-completed decade stirs a slow but seismic shift. The Information Age we all prepared for is ending. Rising in its place is what I call the Conceptual Age, an era in which mastery of abilities that we've often overlooked and undervalued marks the fault line between who gets ahead and who falls behind.

To some of you, this shift - from an economy built on the logical, sequential abilities of the Information Age to an economy built on the inventive, empathic abilities of the Conceptual Age - sounds delightful. "You had me at hello!" I can hear the painters and nurses exulting. But to others, this sounds like a crock. "Prove it!" I hear the programmers and lawyers demanding.

OK. To convince you, I'll explain the reasons for this shift, using the mechanistic language of cause and effect.

The effect: the scales tilting in favor of right brain-style thinking. The causes: Asia, automation, and abundance.

Read on at Wired. Adapted from "A Whole New Mind: Moving from the Information Age to the Conceptual Age", by Daniel H. Pink.

Follow me on Twitter: @IanYorston

Brain has 'early warning system'

BBC News

University College London experts have shown how the brain subconsciously remembers details around past dangers.

Writing in Nature, they say blocking this system could help treat pain by interrupting such a brain process. Researchers said volunteers could not recall details of a test which had led to them getting a mild electric shock, but activity in the brain revealed that they had correctly logged the [associated] data by using a series of complicated computations.

Dr Ben Seymour, who led the research, said: "If we showed a square followed by a circle followed by the painful shock this part of the brain could soon learn to predict that the circle was bad news. After a while, it would learn that the square wasn't that good either, as it was [often] followed by the circle. By recording [and analysing] these chains of events, the brain was able to set early alarm bells ringing in the volunteer."

He added: "The brain is a phenomenal biological computer with around 100 billion nerve cells which determine our thoughts and behaviour. Evolution clearly favours animals that are good at looking after themselves."

Follow me on Twitter: @IanYorston