[Top] [All Lists]

Re: [ontolog-forum] Watson's Wed. win by simple look-up questions

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sun, 20 Feb 2011 08:01:13 -0500
Message-id: <4D611099.7030302@xxxxxxxxxxx>
On 2/20/2011 1:59 AM, doug foxvog wrote:
> Watson's win on Wednesday was made very likely because enough of the
> problems could be solved by simple look-up.  These could have been answered
> by definition look-ups, quotation completion, or searches based on key
> words in the question.  Any question that went beyond these techniques
> was not answered, or not answered correctly by Watson.    (01)

I'll grant that point to a large extent, and I'll add that I
believe much better systems will be coming along soon.    (02)

But I'll also point out that there is no commercially available
question-answering system that can do that kind of lookup from
a question stated in English and answer it with the level of
accuracy and confidence that Watson did.  (On the questions
Watson failed to answer, it showed very little confidence.)    (03)

Cyc can answer some clever questions.  But what could it
do with the range of questions as stated on Jeopardy?
I would very much like to see what Cyc could do, and how
long it would take it to answer those questions.    (04)

And by the way, I'd like to make some comments about those
absurd "human-like" comments by journalists.  The one Leo
cited was truly ridiculous:    (05)

http://www.computerworld.com/s/article/9210319/Brain_behind_IBM_s_Watson_not_unlike_a_human_s?source=CTWNLE_nlt_pm_2011-02-18    (06)

Opening section:    (07)

    "Brain behind IBM's Watson not unlike a human's    (08)

    Like humans, Watson only uses a fraction of its memory to generate
    answers to Jeopardy questions..."    (09)

The opening section of that article is typical journalist-style opinion
about three technical issues that the journalist does not understand:
(1) NLP, (2) human cognition, and (3) any similarities between them.    (010)

If you skip the first few paragraphs of nonsense, the remainder
is factual information about the auxiliary storage used by Watson.    (011)

But I would also like to point out that there are two kinds
of respectable research toward the original goals of artificial
intelligence.  One is the BICA conference on Biologically
Inspired Cognitive Architectures:    (012)

    http://bicasociety.org/2010/bica2010schedule.htm    (013)

I participated in that conference in November 2010.
Following are my slides:    (014)

    http://www.jfsowa.com/talks/bionlp.pdf    (015)

Another series of conferences are devoted to AGI, which is defined
as Artificial *General* Intelligence as opposed to ordinary AI,
which includes special-purpose systems such as Watson or chess
programs that can beat humans on chess, but nothing else:    (016)

    http://agi-conf.org/2011/    (017)

There is overlap between the people who participate in AAAI,
BICA, and AGI.  Some of the AGI participants are overly
optimistic about the chances of implementing intelligence
at the level of humans within the next 20 years.    (018)

I am extremely skeptical about true AGI in that time frame.
But I do believe that we'll see many more systems that can
perform Watson-like NLP on a laptop within the next 10 years.    (019)

John    (020)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (021)

<Prev in Thread] Current Thread [Next in Thread>