Apr 042010
 
Rendering of human brain.
Image via Wikipedia

New Approach Unites Two Prevailing but Often Opposed Strains in Artificial-Intelligence Research

In the 1950s and ’60s, artificial-intelligence researchers saw themselves as trying to uncover the rules of thought. But those rules turned out to be way more complicated than anyone had imagined. Since then, artificial-intelligence (AI) research has come to rely, instead, on probabilities — statistical patterns that computers can learn from large sets of training data.

The probabilistic approach has been responsible for most of the recent progress in artificial intelligence, such as voice recognition systems, or the system that recommends movies to Netflix subscribers. But Noah Goodman, an MIT research scientist whose department is Brain and Cognitive Sciences but whose lab is Computer Science and Artificial Intelligence, thinks that AI gave up too much when it gave up rules. By combining the old rule-based systems with insights from the new probabilistic systems, Goodman has found a way to model thought that could have broad implications for both AI and cognitive science.

Early AI researchers saw thinking as logical inference: if you know that birds can fly and are told that the waxwing is a bird, you can infer that waxwings can fly. One of AI’s first projects was the development of a mathematical language — much like a computer language — in which researchers could encode assertions like “birds can fly” and “waxwings are birds.” If the language was rigorous enough, computer algorithms would be able to comb through assertions written in it and calculate all the logically valid inferences. Once they’d developed such languages, AI researchers started using them to encode lots of commonsense assertions, which they stored in huge databases.

The problem with this approach is, roughly speaking, that not all birds can fly. And among birds that can’t fly, there’s a distinction between a robin in a cage and a robin with a broken wing, and another distinction between any kind of robin and a penguin. The mathematical languages that the early AI researchers developed were flexible enough to represent such conceptual distinctions, but writing down all the distinctions necessary for even the most rudimentary cognitive tasks proved much harder than anticipated.

Embracing uncertainty

In probabilistic AI, by contrast, a computer is fed lots of examples of something — like pictures of birds — and is left to infer, on its own, what those examples have in common. This approach works fairly well with concrete concepts like “bird,” but it has trouble with more abstract concepts — for example, flight, a capacity shared by birds, helicopters, kites and superheroes. You could show a probabilistic system lots of pictures of things in flight, but even if it figured out what they all had in common, it would be very likely to misidentify clouds, or the sun, or the antennas on top of buildings as instances of flight. And even flight is a concrete concept compared to, say, “grammar,” or “motherhood.”

Read more . . .

Reblog this post [with Zemanta]

Other Interesting Posts

  One Response to “Grand Unified Theory of AI”

  1. Embarassingly obvious TOE, expanding the horizon beyond Darwin And Einstein. It is spacedistance, NOT spacetime, that does it:

    Theory Of Everything Without Strings Attached.

    Embarassingly Obvious And Simple.

    See the signature links.

    Life's Genesis Was Not Cells But First Gene's Self Reproduction.

    Life Is Just Another Mass Format.

    Since July 5 1997 I have developed and been proposing the following scenario of life's genesis:

    * Life's genesis was not cell(s), but the self reproduction of yet uncelled ungenomed gene(s).

    * There was NOT any "Pre-History Of Life" evolving in an archaic pre-modern life cell.

    * Cells were definitely NOT life's genesis. Cells were products of evolution of Earth's primal organisms, of Earth's first stratum organisms, the RNA genes that have always been and still are running the show of life, the energy-storing biosphere survival, since Earth life's day one.

    * A gene's self reproduction was distinctly an evolutionary, enhanced energy constraint event, above the earlier, random, radiated-energy-induced genes formations.

    * Every evolutionary step is inherently an event of an enhanced energy constraint.

    * Genomes, RNA and DNA, are functional organs evolved by the primary RNA genes. Cell membranes are also functional organs evolved by the primary RNA gene.

    * Life is but one of the many many mass formats in the universe, and its evolution is driven as the evolution of all cosmic mass formats, to gain temporary enhanced energy constraint, i.e. to survive as long as possible.

    Dov Henis

    (Comments From The 22nd Century)

    03.2010 Updated Life Manifest
    http://www.the-scientist.com/community/posts/list
    Cosmic Evolution Simplified
    http://www.the-scientist.com/community/posts/list
    Gravity Is The Monotheism Of The Cosmos
    http://www.the-scientist.com/community/posts/list

    PS:

    This Theory Of Everything, with definition of evolution, covers also ALL aspects of anthropology. DH

    =============

    TOE: Religion Or Science?

    (Fwd from the-scientist.com:)

    I.

    [quote=BobTS1162939] This is the Theory Of Everything In A Nutshell (TOEIANS):

    Basic construction of the universe: 1. Particles 2. Strings 3. Frames.

    Each particle has string. They combine with each other into quantum and physical objects.

    Particles' travel along their strings appear to us as: 1. Gravity. 2. Properties. 3. Forces.

    String-particle travels/lives within a frame seen as: 1. Waves. 2. Feelings 3. Influence.

    Time is a one dimensional string whose constant value is 9. The universe is constructed on the number 3. Time moves outward dragging space with it. This outward expansion causes space which is a string to distort / stretch which we see as repairs / deterioration / aging / etc.. If the strings were decreasing we would see the reverse. Things would essentially become younger until they simply disappeared as opposing to dying / being not repairable as we see / experience now. The development of language and the effort to define things means particles / strings / frames have different names depending on the discipline.[/quote]

    II. My comment

    A) Since Life is, by our sensory conception, a virtual reality affair, religion is a legitimate virtual reality tool for going through life. But I am not religious. My senses do not become affected by the above TOEIANS. I embarassingly admit that hard as I try I am unable to comprehend the above TOEIANS.

    B) My own conception of TOE is scientific, not religious, based strictly on data recorded and observed, of ubiquitous cosmic phenomena. And in presenting my TOE conception I do not deal with mechanisms but with the base processes.

    Dov Henis

    (Comments From The 22nd Century)

Leave a Reply

%d bloggers like this: