markram

medium.com
The boy whose brain could unlock autism

SOMETHING WAS WRONG with Kai Markram. At five days old, he seemed like an unusually alert baby, picking his head up and …

 “There are elements of the intense world theory that better match up with autistic experience than most of the previously discussed theories,” says Ari Ne’eman, president of the Autistic Self Advocacy Network, “The fact that there’s more emphasis on sensory issues is very true to life.”

[…]

“We agree that autistic people do have a number of cognitive advantages and it’s valuable to do research on that,” he says. But, he stresses, “People have worth regardless of whether they have special abilities. If society accepts us only because we can do cool things every so often, we’re not exactly accepted.”

Artificial Brains

 

External image

At what point will a program make a break from its human origins and become something independent and just plain weird? Intellectual property makes assumptions about human authorship and fixation that are becoming obsolete with new computing models. While self-aware artificially intelligent robots may still be pretty far off, there are some very strange things going on in experimental computing.

As Dorothy might say, “I’ve a feeling we’re not in Kansas anymore.”
 
  Artificial intelligence researchers have posited different perspectives on what it means to be “creative.” In many ways, it involves the ability to take in input and process it in a way that results in a novel combination of pre-existing ideas and information. Some forms of creative thought are more innovative and human than others. For example, once a given conceptual space is established an entity may process pre-existing ideas in a creative way, evidencing “exploratory creativity.” On the other hand, higher or “transformational creativity” is achieved when the conceptual input is transformed by rejecting or redefining previous conceptual constraints to produce creative output that transforms our understanding, raising it to a new level. Watson (IBM’s AI Jeopardy! champion) showed a high degree of exploratory creativity by being able to process massive amounts of cultural information in a human-like way. It had to zero in on a few essential key words, executed thousands of language analysis algorithms in parallel, then would buzz in when the possibilities had been narrowed down to an answer for which it was sufficiently “confident.” Also interesting is that when Watson’s answer was off, it tended to be hilariously off, which reflects the difficulty of sorting associations in a human-like way.
 
In 1950 Alan Turing proposed the “Turing test” to test a machine’s ability to appear human. Participants would converse with the machine or a human in a text-only format. They would then indicate if they believed they were communicating with a human or with a machine. The machine would pass the test if it could generate answers that were indistinguishable from a real human. Essentially, a program may be considered somewhat artificially intelligent if it seemed like a human to other humans. Watson may be considered a highly developed example of artificial intelligence according to his mastery of human-like thought and language. Yet while this is an impressive feat of computing, Watson did not in fact learn or rewrite its underlying programming or evidence any degree of transformational creativity. This does show that super-computers are getting powerful enough to process information quickly enough to result in human-like output in response to input.

Neurologists are in the process of reverse-engineering a biologically accurate brain down to the molecular level using a supercomputer. Headed by Henry Markram, the Blue Brain Project is programming elements of the supercomputer to model individual neurons, building to the various brain regions and intends to model the complete human brain within ten years. This is particularly tricky and computation-intensive due to all the interconnection and feedback loops within a human brain, in direct contrast to input-output models which are the industry standard.

I could go into neocortical columns and interconnected influence of one brain region on another, but more interestingly is the question: if a simulated brain DID behave just like a human one, would it be self-aware? (this is a legal mess which I may deliberate upon some other day) Computers are being built which are able to act and literally think like humans. Some very interesting developments in programming are behind these developments. In addition, some computing methods which do not mimic human action at all behave in a way that calls into question legal assumptions about creativity and human authorship.   

In conclusion - some more from the Wizard of Oz:

“Scarecrow: I haven’t got a brain… only straw.

Wizard of Oz: Why, anybody can have a brain. That’s a very mediocre commodity. Every pusillanimous creature that crawls on the Earth or slinks through slimy seas has a brain. Back where I come from, we have universities, seats of great learning, where men go to become great thinkers. And when they come out, they think deep thoughts and with no more brains than you have. But they have one thing you haven’t got: a diploma.”