[ExI] watson on jeopardy

Darren Greer darren.greer3 at gmail.com
Thu Feb 17 21:29:39 UTC 2011


>Watson passes or is getting close in Jeopardy, neither can pass at general
language.<

Yes, I meant, and I guess Kurzweil meant, that Watson is one step closer to
passing the general language test though he and computers in general still
maybe be very far away. Because it isn't just a matter of the computer
finding the right answer to a direct question -- he has to place the words
in context first and then hunt for the answer.  The example given by one of
the programmers was if someone 'runs' down the street and someone else
'runs' for president, Watson has to be able to sort out  which meaning is
intended before he can begin to word associate. And this ability moves
Watson closer to passing the general language test than a computer has ever
been before, does it not?

I think one of the more interesting aspects of this Watson discussion is
that we as a group are very focussed on where we want to be as opposed to
where we actually are. I think it also is interesting that many of the
things Watson does we do as thinking machines as well. We don't understand
the programming, or the platform. And we have only the basest understanding
of the hardware. But I know when presented with a question I put the words
in context and my brain begins to associate. Other than thinking in image as
well as language and numbers, I don't understand how Watson, in this one
area at least, is radically different from me. Except he has a far greater
knowledge base of raw trivia stored in accessible 'cells.'





d.

2011/2/17 spike <spike66 at att.net>

>
>
>
>
> *…* *On Behalf Of *Darren Greer
> *…*
>
>
>
> I just read the Kurzweil article and he points out that Watson is much
> closer to being able to pass the Turing test than a chess playing computer
> as it is dealing with human language. And so based on that criteria, it is a
> step forward no matter how you slice it.
>
>
>
> d.
>
>
>
> Chess programs have already passed the Turing test in chess, a long time
> ago.  So Rybka wins the Turing test at chess, Watson passes or is getting
> close in Jeopardy, neither can pass at general language.
>
> spike
>
>
>
>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
>


-- 
*There is no history, only biography.*
*
*
*-Ralph Waldo Emerson
*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20110217/45bddd19/attachment.html>


More information about the extropy-chat mailing list