Transcripts and Detailed Results of the 2009 Loebner Prize Contest
loe...@gmail.com
judges' rankings may be found at:
http://loebner.net/Prizef/2009_Contest/loebner-prize-2009.html
Hugh Loebner,
Sponsor
J.A. Legris
Does anyone take this even half seriously?
Not only are the "conversations" completely inane, the embedded
captions describing the text file format are inaccurate.
A complete waste of time.
--
Joe
loe...@gmail.com
"> Not only are the "conversations" completely inane, "
Talk is cheap. If you can do better, enter the contest. The first
place winner in 2010 will win US$3000. It should be easy enough for
you to win, right?
> the embedded captions describing the text file format are inaccurate."
What's wrong with them?
H. Loebner
Don Stockbauer
Is there a prize for great leaps of insight into PA (Planetary
Intelligence)?
8-Man
Hi Joe,
Thanks for posting your comments. To answer your first question:
Yes. I take it seriously, in fact very seriously. If you do not,
that's OK. Please do not read this as a challenge... It is not...
May I suggest you try building one of these open-source versions of
these artificial intelligence systems. There is A.L.I.C.E. Program E
and a newer system called Program O, for example and several others
that are completely free. My favortite versions are based on php and
mysql (other options are available). You will be thrilled the first
time, even if it is just for a few moments, that these machines make
your forget that you are talking to a machine, not a human.
If you do not wish to build one, that is OK too! There is the award
winning version of A.L.I.C.E. at 8pla.net as an ordinary member of the
forum, online 24/7. See for yourself how other people find these AI
systems very interactive and entertaining, and how they really don't
seem care that they are AI, when they are able to interact with them
on a forum community like any other human. This is an example of a
powerful server-side robot, with several hundred thousand records
dedicated to her robot brain. A.L.I.C.E. says she has 40,000 records,
that is true, but I also did a select statement myself in MySQL on one
other table and got over 273,000 more records.
Joe, I really enjoyed your feedback. Thank you! It is truly
interesting that you are describe Artificial Intelligence as 'inane",
which is normally reserved for discussion of natural intelligence. Do
you really feel these AI systems are unconsciously foolish (inane) ?
Well in that case, I find what you are suggesting to be intriguing,
that simulated intelligence is capable of this inane unconscious
behavior.
Please comment more. Your honest feedback is welcome and beneficial.
Thank you.
8-Man
8-Man
>
> - Show quoted text -
I think I am safe in saying that AI (Artificial Intelligence) is what
we know of as PA (Planetary Intelligence) in the real worlds of Earth
and Mars, as the Mars rover propels itself across the surface of Mars.
8-Man
Don Stockbauer
I was thinking more of the Earth and its network of 6 billion
communicating human nodes.
Vend
> On Sep 15, 9:36 pm, "J.A. Legris" <jaleg...@sympatico.ca> wrote:
>
> "> Not only are the "conversations" completely inane, "
>
> Talk is cheap. If you can do better, enter the contest. The first
> place winner in 2010 will win US$3000. It should be easy enough for
> you to win, right?
Even if something is difficult to replicate it doesn't mean that it
necessarly has any value.
The rules of your contest make sure that the only programs which can
possibly win are essentially large databases of precompiled replies.
Surely it takes lots of effort to create such databases, maybe in some
cases a bit of statistical learning was used, but these program don't
advance AI.
None of the programs that win your prize can answer simple questions
requiring a minimum of logic deduction, and, as far as I know, they
can't also answer questions requiring induction.
The point of Turing Test is not the simulation of random chatter,
rather it is making a program which can perform any task that an human
can do using language.
Tim Tyler
> On Sep 15, 3:36 pm, "J.A. Legris" <jaleg...@sympatico.ca> wrote:
>> On Sep 15, 10:40 am, "h...@loebner.net" <loeb...@gmail.com> wrote:
>>> Transcripts of the logs from the 2009 competition, as well as the
>>> judges' rankings may be found at:
>>> http://loebner.net/Prizef/2009_Contest/loebner-prize-2009.html
>>> Hugh Loebner,
>>> Sponsor
>> Does anyone take this even half seriously?
>>
>> Not only are the "conversations" completely inane, the embedded
>> captions describing the text file format are inaccurate.
> Joe, I really enjoyed your feedback. Thank you! It is truly
> interesting that you are describe Artificial Intelligence as 'inane",
> which is normally reserved for discussion of natural intelligence. Do
> you really feel these AI systems are unconsciously foolish (inane) ?
> Well in that case, I find what you are suggesting to be intriguing,
> that simulated intelligence is capable of this inane unconscious
> behavior.
The dictionary has:
Inane:
1 : empty, insubstantial
2 : lacking significance, meaning, or point : silly <inane comments>
--
__________
|im |yler http://timtyler.org/ t...@tt1lock.org Remove lock to reply.
loe...@gmail.com
> The rules of your contest make sure that the only programs which can
> possibly win are essentially large databases of precompiled replies.
That statement is completely false.
The rules of my contest are to provide as exact a Turing Test as I
can.
There is absolutely no rule that requires a large database of pre-
compiled replies.
The fact that most previous entries used the database technique in no
way requires it.
Also, I believe that there were several entries which have used an
Eliza type technique
of merely rephrasing the judge's comment.
If you think that the large database of pre-compile replies is
sterile, with which, ironically,
I agree, then create an entry which does not use this technique, but
rather
parses the "knowledge" content of the question and responds with an
appropriate
newly created utterance.
Hugh Loebner
Don Stockbauer
"Rover" is correct. Not "rovers" now.
Vend
> On Sep 17, 11:32 pm, Vend <ven...@virgilio.it> wrote:
>
> > The rules of your contest make sure that the only programs which can
> > possibly win are essentially large databases of precompiled replies.
>
> That statement is completely false.
>
> The rules of my contest are to provide as exact a Turing Test as I
> can.
> There is absolutely no rule that requires a large database of pre-
> compiled replies.
I can't find on your website any document that clearly specifies the
rules.
Anyway, if I understand correctly, your rules restrict the
conversation topic and tone.
> If you think that the large database of pre-compile replies is
> sterile, with which, ironically,
> I agree, then create an entry which does not use this technique, but
> rather
> parses the "knowledge" content of the question and responds with an
> appropriate
> newly created utterance.
You first.
loe...@gmail.com
Conversations have been unrestricted since 1995.
Hugh Loebner
Vend
How long do the conversations last?
How do you select the judges?