Re: Why is NLP such a hard topic?

My 'take' on this is largely that we largely try and do (fit) NLP into
a computational framework that just isn't right (suited to the REAL
problem) - it's like the chess computer vs. the human - we can build a
decent chess computer these days BECAUSE the rules of chess fit
reasonably well with what a computer can do.  However, as we all know,
a human player plays chess very differently.

So, to do NLP (it seems to me) we ultimately need a machine that
mimics a human's mind (it's not a case of clock-cycles, it's one of
architecture) - then that 'mind' can be taught at computational
speeds.  Once we've done that (ahem!) we can all go and look for
alternative jobs!

A related question - and something I seem to remember being raised my
Marvin (Minsky - not the robot) - do you think a computer with such a
brain would be able to cope with its own reality ... if we make it
human-like, won't it go mad 'knowing' it's actually a machine?

A good read for anyone interested in NLP and machine intelligence:
HAL's Legacy (MIT Press)

peetm



On Thu, 6 Nov 2003 09:24:17 +0100, "Niki Estner"
<[EMAIL PROTECTED]> wrote:

>I have long been wandering about that question, and did not really find any
>good answers:
>There are many (profitalbe) uses for natural language processing (automated
>translation, search engines, spam filters, database querys...), and lexicons
>(even including kind of semantic information) are available, and yet (as far
>as I know) there is no really good implementation for natural language
>processing (Think of something like English in BNF).
>
>Why is this the case?
>
>It's surely not because it would be too much work to write down a few
>thousand rules or more (theres money in it).
>
>Can anyone explain me (or point me to information that does) why this is
>such a hard topic?
>Is the number of rules simply too big???
>Can syntax not be analyzed without semantics?
>Is some mystical "soul" required to process natural language????
>
>Thanks a lot
>
>Niki
>