I'd like to jump into the debate between Keith Sutherland and Pat Hayes over a couple of points, particularly as my views have been quoted.
Pat Hayes said:
Theres a persistent misunderstanding lurking here: that a program is like a recipe which the computer 'obeys' blindly. Some are like that, but others arent. Some programs are more like populations which evolve, some work by performing inferences in organised ways, some do operations which look like shufflings and sortings of data, some lie in wait and pounce when patterns appear, some of them adapt and learn, some of them send themselves from place to place around the world leaving special trails in networks. They do all kinds of things in all kinds of machine architectures. ...
Well I think some of this is giving unnecessary colour by the words chosen. I do not believe that programs 'learn', 'adapt', 'pounce' and so on, except by very broad analogy to how humans do these things. Programs mostly implement only three things (sequence, selection and iteration) by the way in which they can be pushed through a central processing unit(CPU) which only accepts inputs sequentially. Yes you can have multiple CPUs and parallel processing which means you can do lots of things sequentially though I think this merely means you do a lot of not-very-clever things in parallel. My argument is that all the baggage about non-procedural programming, neural networks, genetic algorithms etc operates purely at the conceptual level but whether your high-level language is COBOL, a word-processing package or a genetic algorithm, the low-level implementation uses the same lowly sequential CPU. This is what Keith Sutherland referred to when he said:
There have been many attempts to argue that contemporary models like neural networks and genetic algorithms operate along more "biological" principles. Again, I'm not qualified to comment, but would refer to Sam Salt's Tucson presentation (1996) where he pointed out that the differences are exaggerated -- all programming models ultimately reduce to a series of binary digits going through a Turing Machine.Pat then said:
Well, replace 'Turing Machine' with 'hardware' (a Turing machine can't do anything, its just a mathematical abstraction), and I'll agree. But this is like saying that all brains ultimately reduce to masses and masses of quarks and leptons exchanging photons: true, but not very useful. This doesnt account for how programmed machines work, how they do what they do.He's right - a Turing machine is an abstract mathematical concept. However, I think most computer scientists will agree that a standard computer is pretty much an instantiation of the idea of a universal Turing machine - I was being careless with my words. In my presentation I was trying to draw attention to the fact that there is a strong contrast between the conception of programs such as artificial neural network programs and the mundane implementation in a computer CPU. The practice of giving programs names such as genetic algorithms etc seems rather more to do with marketing and trying to make it sound good than the reality of what they really do. It seems rather like the anthropomorphising of animals.
Later Keith says...
One is tempted to say that this particular road to Damascus was just the result of the realization that GOFAI was a blind alley....and Pat replies:
Well, this 'blind alley' still has an amazing amount of traffic on it. I wonder where you get this impression that GOFAI, as it is affectionatly known, has somehow died or gotten refuted? Let me invite you to the next AAAI or IJCAI or ECAI or IAAI meeting to get a better impression about what is going on.On the whole I have to say I agree with Pat on this. Again, in my Tucson presentation I made a comment that "GOFAI is dead, long live Artificial Neural Networks" but I was trying to be be ironic (sarcastic?). I do thing GOFAI has achieved great things and still has lots to offer; ditto neural networks. My main argument is that we shouldn't get too excited about these techniques when it comes to consciousness research. As technologies, both GOFAI and Neural Networks are very useful - I just don't think they have much meaning when we try to apply them to human consciousness.
Sam Salt
(d.w.salt@derby.ac.uk)
jcs-online menu