Computers designing chips that humans don't understand ...

Discussion in 'Trading Software' started by aphexcoil, Jan 19, 2003.

  1. Scientific American -- Recent Article

    (Uh-oh! They're evolving!)

    Evolution is an immensely powerful creative process. From the intricate biochemistry of individual cells to the elaborate structure of the human brain, it has produced wonders of unimaginable complexity. Evolution achieves these feats with a few simple processes--mutation, sexual recombination and natural selection--which it iterates for many generations. Now computer programmers are harnessing software versions of these same processes to achieve machine intelligence. Called genetic programming, this technique has designed computer programs and electronic circuits that perform specified functions.
  2. So when will we see "Mother" like the movie Metropolis?

    Or Terminator?



    Anyways, interesting stuff... thanks Aphie...
  3. Actually, I suspect that even traditionally engineered chips are already far beyond human understanding. For example, simulating the behavior of a pentium chip creates about 60 megabytes of data per clock cycle. The only way to design chips like CPUs, GPUs and high speed networking chips is to completely break them down into little pieces and use large teams of engineers on each little piece. Nobody can fully understand all of the workings of complex chips (or complex software systems like a modern OS).

    The use of automated optimization techniques, such as those based on genetic algorithms, only makes the chips that much more opaque in their functioning. But do we really need to know why they work as long as we know that they work (or more importantly that we know of a process that make them work better).

    All of this leads to the topic of distributed cognition -- where a social system is "smarter" than any of the individual people (or agents) in it. When things get complex, nobody has the full picture. Yet if the system is well designed, it will work even when most of the members are largely ignorant of most of the details.

    If you think genetic algorithms can evolve interesting solutions to tough problems, wait until somebody creates a socially cooperative set of genetic algorithms. That will be very interesting.

  4. You will be assimilated. Resistance is futile. Individuality is irrelevant.
  5. (You have to subscribe to Scientific American) ...

    "THE TWO CIRCUITS shown below are both cubic signal generators. The upper circuit is a patented circuit designed by a human; the green and purple parts of the lower circuit were evolved by genetic programming (the other parts are standard input and output stages). The evolved circuit performs with better accuracy than the humandesigned one, but how it functions is not understood. The evolved circuit is clearly more complicated but also contains redundant parts, such as the purple transistor, that contribute nothing to its functioning."

  6. EVOLUTIONARY PROGRAMMING, originally conceived by Lawrence J. Fogel
    in 1960, is a stochastic OPTIMIZATION strategy similar to GENETIC
    ALGORITHMs, but instead places emphasis on the behavioral linkage
    between PARENTs and their OFFSPRING, rather than seeking to emulate
    specific GENETIC OPERATORs as observed in nature. Evolutionary
    programming is similar to EVOLUTION STRATEGIEs, although the two
    approaches developed independently (see below).

    Like both ES and GAs, EP is a useful method of optimization when
    other techniques such as gradient descent or direct, analytical
    discovery are not possible. Combinatoric and real-valued FUNCTION
    OPTIMIZATION in which the optimization surface or FITNESS landscape
    is "rugged", possessing many locally optimal solutions, are well
    suited for evolutionary programming.

  7. The GENETIC ALGORITHM is a model of machine learning which derives
    its behavior from a metaphor of some of the mechanisms of EVOLUTION
    in nature. This is done by the creation within a machine of a
    POPULATION of INDIVIDUALs represented by CHROMOSOMEs, in essence a
    set of character strings that are analogous to the base-4 chromosomes
    that we see in our own DNA. The individuals in the population then
    go through a process of simulated "evolution".

    Genetic algorithms are used for a number of different application
    areas. An example of this would be multidimensional OPTIMIZATION
    problems in which the character string of the chromosome can be used
    to encode the values for the different parameters being optimized.

    In practice, therefore, we can implement this genetic model of
    computation by having arrays of bits or characters to represent the
    chromosomes. Simple bit manipulation operations allow the
    implementation of CROSSOVER, MUTATION and other operations. Although
    a substantial amount of research has been performed on variable-
    length strings and other structures, the majority of work with
    genetic algorithms is focussed on fixed-length character strings. We
    should focus on both this aspect of fixed-lengthness and the need to
    encode the representation of the solution being sought as a character
    string, since these are crucial aspects that distinguish GENETIC
    PROGRAMMING, which does not have a fixed length representation and
    there is typically no encoding of the problem.

  8. Sealed within a transparent, tapered, liquid-filled cylinder, illuminated colored globs slowly rise and fall. Meandering and deforming, their shapes and paths change unpredictably. Invented in 1963, a decorative fixture in many homes during the 1970s, and still in production, Lava Lite lamps are now the object of renewed curiosity.

    Indeed, researchers have come up with a novel application of the mesmerizing movements of the lamp’s globules. They use them as the starting point for generating a sequence of random numbers. Called Lavarand, the random-number generator is the tongue-in-cheek work of Robert G. Mende Jr., Landon Curt Noll, and Sanjeev Sisodiya of Silicon Graphics in Mountain View, Calif.

    Random numbers are an immensely valuable commodity, not only for the operation of computer-based slot machines but also for computer simulations and for generating the secret strings of digits required to encode and decode sensitive information in cryptographic systems.

  9. Without taking one side of Jaron's dogma or another (place me somewhere else entirely) I would disagree strongly with his "Argument from Software" — which is as flawed as Bishop Wilberforce's Argument from Design.

    Back in the days when programs could be debugged but processing could not be counted on from one kilocycle to the next, John von Neumann wrote his final paper in computer theory: "Probabilistic Logics and the Synthesis of Reliable Organisms from Unreliable Components" [in Claude Shannon and John McCarthy, eds., Automata Studies (1956) pp. 43 — 99]. It makes no difference whether you have reliable code running on lousy hardware, or lousy code running on reliable hardware. Same results.

    What should reassure the technophiles, and unsettle the technophobes, is our world of lousy code. Because it is lousy code that is bringing the digital universe to life, rather than leaving us stuck in some programmed, deterministic universe devoid of life. It is that primordial soup of archaic subroutines, ambiguous DLL's, crashing Windows, and living — fossil operating systems that is driving the push towards the sort of fault embracing template — based addressing that proved so successful in molecular biology, with us — and our computers — as one of its strangest results.
  10. You must br wild not aphie!!!

    You paste and post too much!!!

    Must warn Aphie, Wild hacked in Aphie's Supercomputer...
    #10     Jan 19, 2003