|
[DPRG] Help with Elementary School Project
Subject: [DPRG] Help with Elementary School Project
From: R. Steven Rainwater
srainwater at ncc.com
Date: Wed Apr 25 12:00:11 CDT 2007
On Sun, 2007-04-22 at 15:24, Winefred Washington wrote:
> If possible, could the members of this list take time to reply
> to a few of his questions.
I just saw this - okay, I'll take a shot at it. A lot of these questions
seem a bit advanced for elementary school but it's cool that they're
thinking about this stuff. Hopefully I don't get points off for grammer
and spelling!
> 1. How do autonomous robots collect information and make
> decisions off of it?
Just like biological animals evolved special structures such as eyes and
ears for sensing the world, autonomous robots can be equipped with
electronic sensors that can see or hear. More frequently they have very
simple sensors that detect walls or nearby obstacles. Information from
the sensors is sent electronically to a computer that acts as the
robot's brain. A program, usually written by a human, generates
behaviors based on the sensor inputs.
> 2. What types of autonomous robots do you make?
I mostly seem to make robots that never quite get finished. At present
I'm working on a small, behavior-based robot. It has two rubber tracks
instead of wheels and has differential drive, meaning that it can turn
in place. It uses a Motorola 68332 computer (almost an antique!) to run
software written in C.
> 3. How do these autonomous robots help us?
My robots help me not have any spare time and prevent the unnecessary
accumulation of spare cash.
Most other autonomous robots help researchers or hobbyists learn more
about how other autonomous creatures, such as ourselves, might work.
Commercially available autonomous robots like the iRobot Roomba help
clean floors.
> 4. How do you make your autonomous robots?
To be autonomous, a robot must be capable of directing its own actions.
The words autonomous and automated are often confused. Automated agents
perform pre-programmed actions without intervention but are not
self-directed. Autonomous agents are self-directed and do not require
any other agent, such as a human, to tell them what to do and when to do
it. You and I are autonomous, not automated. A factory assembly line is
automated but not autonomous. Many robots are automated and simply
execute pre-programmed actions. However, many others are autonomous
robots.
Because researchers don't fully understand how biological animals work,
they are not yet sure of the best way to make a robot autonomous. One
type of autonomous robot that is well understood is called a
behavior-based robot. These often use a mechanism called subsumption to
provide the autonomy. This has proven to be a very successful method but
is limited to duplicating the simple level of autonomy experienced by
single celled animals and insects. The iRobot Roomba, research robots,
and many hobbyist robots use this method to achieve autonomy.
> 5. How do you make the information the robots start with?
I'm not sure what the question means. I'll assume it refers to the
software or program within the robot. There are several methods of
generating software for a robot but the most common is for a human to
write it in a programming language. Two other methods are to let it
evolve using software that simulates biological genetic evolution, or to
use software that simulates neural networks to learn from test data.
> 6. What is the difference between free will and autonomous?
They never talked about philosophy at my elementary school.
This is mostly a "semantic" issue - that means any time you hear people
talking about this subject, they're mostly arguing about the definition
of the words and not the underlying question. It also means you'll never
get an easy, straight answer because everyone you ask will tell you
something different. Both terms suggest that the agent has the ability
to self-direct its actions.
Autonomy is a more technical term with a more restricted meaning that
would likely be understood and accepted by most people. You could call
yourself or your robot autonomous and not create any controversy.
Free Will has no agreed upon definition and means something different to
each person. It may have religous or supernatural meanings to some and
not to others. Some philosophers insist it describes something that
doesn't exist while others insist it does. If you claim your robot has
free will, almost everyone will object. If you claim you have free will,
about half the people who hear you will object.
The difference is usually the source of the self-direction. Very old
definitions of free will required the source of self-direction to be
outside the natural world, in a soul or some other source that can't be
measured or detected. In general, this idea is called dualism. Modern
dualists have switched from souls to ideas such as quantum
spin-mediation as the other-worldly, unknowable source of free will. The
opposing opinion is known as materialism and holds that whatever free
will is, if it exists, it comes from something natural going in the
animal.
Generally dualists believe that only agents made from meat, from
biological stuff, can be alive, conscious, and have freewill (because
non-meat-based agents can't reproduce the otherworldy connection to a
free-will source). They believe building robots with these qualities is
fundamentally not possible.
Materialists, on the other hand, usually believe that by learning how
consciousness and free will occur in animals, we can learn to duplicate
them algorithmically. Materialists usually believe that robots with
consciousness and free will are possible but may be very complex and
beyond our current technological reach.
Both a materialist and a dualist would probably accept that autonomous
robots were possible.
> 7. Do you think robots and animals deserve the same rights?
There is no general agreement today on what rights animals deserve. For
the animal called human, there's a little bit of agreement. Many people
believe rights are "social constructs", that is, agreements between a
group and an individual, others believe rights are "natural" by which
the usually mean granted by a supreme being and inalienable. Also, some
groups use the words "animal rights" when they really mean something
like "humans should be nice to other animals".
If rights are social constructs, non-human biological animals can't have
them because they can't understand or agree to the social contract. If
rights are natural, it depends on which supreme being you think granted
the rights and whether or not he/she/it provided a list of which rights
the animals got. If animals having rights just means humans should be
nice to them, most humans agree, at least when they're not eating
animals or being eaten by them.
All existing robots are out of luck. I'm not aware of any religion which
mentions them (with the possible exception of Shinto if you buy the
claimed connection between Karakuri Ningyo and modern robots) and no
autonomous robots have reached the level of being able to understand and
enter into social contracts of their own "free will".
If robots reach the same level of autonomy and intelligence that humans
posses, they could claim equal rights under social contract theory. They
could also make a plausible claim under natural rights theory, at least
within several religious traditions. The UK is currently studying the
social implications of sentient robots demanding "human" rights. In
other countries such as South Korea, there seem to be early moves in
progress to legally classify robots in a way that precludes them gaining
"human" rights.
Hmmm... I never actually answer that one did I? Okay, for appropriate
definitions of robots, animals, and rights; yes, I believe they should
have the same rights; otherwise, no. :)
-Steve
More information about the DPRG mailing list
|
