---------- Forwarded message ----------
Date: Sat, 9 Sep 2000 16:01:27 -0700 (PDT)
From: Franklin Wayne Poley <fwpoley@vcn.bc.ca>
Reply-To: Robot-for-President@egroups.com
To: robot-for-president@egroups.com
Subject: [Robot-for-President] [EDTV-Robotics-State-Of-The-Art] Re: Can
HomeMate Surpass Human Equivalency in Measured Intelligence? (fwd)
---------- Forwarded message ----------
Date: Sat, 9 Sep 2000 16:00:39 -0700 (PDT)
From: Franklin Wayne Poley <fwpoley@vcn.bc.ca>
Reply-To: EDTV-Robotics-State-Of-The-Art@egroups.com
To: "Shepard, Prue" <Shepard@helpmate.com>
Cc: Marvin Minsky <minsky@media.mit.edu>,
edtv-robotics-state-of-the-art@egroups.com, jfe@helpmate.com,
andrew.kun@unh.edu, joanna@ai.mit.edu, humanoids@usc.edu,
hans.moravec@rover.ri.cmu.edu
Subject: [EDTV-Robotics-State-Of-The-Art] Re: Can HomeMate Surpass Human
Equivalency in Measured Intelligence?
On Thu, 31 Aug 2000, Shepard, Prue wrote:
> This is Prue sending for Mr. Engelberger.
>
> Marvin,
>
> I was not "flimflamming" any one! I have long defended your views on the
> elusiveness of "human equivalency." And, I never promised you a rose garden.
Human equivalency is not elusive for hundreds or thousands of practical
everyday criteria of performance. When mobile humanoid robots are on the
mass market I GUARANTEE you that they will be evaluated by human
equivalency criteria. HomeMate is supposed to simulate some everyday
household tasks. Why buy one if it doesn't at least approximate human
equivalency or substitute for a human worker?
> What I promise is a personal robot of great utility to a cognitive but
> physically impaired individual. The robot would extend the period of
> independent living at a monthly cost of a luxury automobile rental.
Good enough. But let's find out what the most luxurious model money can
buy might be like, ie Robotics-State-Of-The-Art.
> I am appending my Gauntlet article sans illustrations. This is about
> robotics, not AI. The HomeMate would not address any of the challenges of
> learning or verbalization or invention or consciousness. Lenat speaks to AI;
> well said, but of no concern to a roboticist of the here and now.
It is of concern to me as a prospective BUYER. "Optional extras" eh?
> Hans Moravec cleverly predicts 50 years out - when he probably will not be
> around to face his critics.
>
> Franklin's question, "Can HomeMate surpass Human Equivalency in Measured
> Intelligence?" must be answered with a resounding NO. Not now, and maybe
> never.
The profiling exercise I just went through PROVES to you that humanoids
can surpass humans on a number of criteria on the profiles of real human
intelligence vs. artificial humanoid intelligence.
> Utility does not require human equivalency.
There must be a satisfactory simulation of human work performance. I call
that a human equivalency criterion.
The robotics secret is apriori
> knowledge and rules. We tell HelpMate what and where everything is in a
> hospital and we give it 1500 rules to cope with surprises. It is not a
> Rodney Brooks' creature banging around and trying to learn about its
> environment.
Some might even call those 1500 rules an operationalization of common
sense.
> I defy anyone to contest the ability of my speculative HomeMate to perform
> all of the tasks that I propose. And, by the way, they do not include any of
> the cute Minsky chores that are surely beyond current technology.
Ah come on...why not call it the "Minsky Model I"?
> Crucial to all this is a magnificent physical execution. You cannot make a
> silk purse from a sow's ear. No amount of software could save a Hero robot.
> Incidentally, when Jim Crowley saw my presentation on the prototype
> two-armed, mobile, sensate, articulate research robot we built for NASA, he
> exclaimed, "that's a platform to die for."
Good to hear. Congratulations!
And it would be a great platform
> for researchers to work with to explore avant garde concepts. The papers I
> read all seem to oppress PhD candidates with junk hardware.
That's what we need. SOMETHING on the mass market which lots of people
from consumers to students to pros can work on. Something good to start
out which we all know can and will be improved. I think you've got it.
> Finally, since I copy in Franklin, I admonish him to go easy on predicting
> human equivalency or superiority. Even with his premise of an unlimited
> budget, that may not be a realistic or even a worthy goal. My late friend
> Isaac Asimov never saw the need for a robot to be equivalent to a human.
> Robots could be much less then as well as much more than humans.
>
> Joe
>
> <<Gauntlet for Elder Care.doc>>
I am not talking about creating a Frankinstein-like human. I am only
saying that human equivalency is a very common sense and necessary
consideration when it is defined as I have defined it. If a machine can't
do a job as well as a human (more or less) why build the machine?
FWP
To unsubscribe from this group, send an email to:
EDTV-Robotics-State-Of-The-Art-unsubscribe@egroups.com
*** The Era of Total Automation is Now ***
This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 17:37:35 MDT