IUBio

Robot Dreams

George Bajszar gyuri at usa.net
Sun Sep 10 19:33:10 EST 2000


In article <8pg9rm$k68$1 at nnrp1.deja.com>,
  mentifex at scn.org wrote:
> For a robot to have a dream, its AI mind software must shut down the
> input sensorium of strong sensations coming in from the outside world
> or from a virtual world.  In humans, the same process is known as
> falling asleep.  Then the AI software must permit minor "brainstorms"
> of free associations in the mind of the sleeping robot.
>
> If a kind of self-sustaining "weather-pattern" of internal association
> develops, the activity reestablishes a non-waking consciousness in the
> mind of the sleeper.
>
> Because the robot dream is happening after all inside a computer,
> all robot dreams may be monitored as images and sounds in a kind of
> theater of the unconsciousness for analysis or even for reentry by
> the robot into one of its former dreams.  In other words, robots may
> have a much more active dream life than humans do, with such robot
> options as re-experiencing dreams or even of sharing dreams in the
> same dream state with other robots.  However, such co-dreaming by
> robots would be a human programmer's nightmare, because the
> associative vortex in one robot mind would have to be coordinated
> with an identical or at least similar associative vortex in another
> robot mind.  It may even be necessary for a master-slave relationship
> to be agreed upon before two robots can share a mutual dream,
> so that the shared consciousness will flow freely under the
> associative direction of one unitary mind at a time.
> Of course, two robots could take turns in directing the stream of
> consciousness central to the experience of the dream, with a more
> experienced robot digging up a greater variety of old memories
> and transferring the memories as shared mutual dream content to
> a perhaps younger, less experienced robot.  However, there could
> not be too great a gap in the levels of experience between the two
> robots sharing a dream, or the junior robot may not have the
conceptual
> and epistemological wherewithal to absorb the conceptual constructs
> being transferred from the more advanced mind to the neophyte mind.
>
> --
> http://www.geocities.com/Athens/Agora/7256/dreams.html : AI Dreams
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>

There are reasons people sleep. We know if we don't get enough sleep,
our short and long term memory is stronly affected, so it is a natural
conclusion to think that a cleanup or reorganization of our memory is
occurring during sleep, more particularly the dreaming process is
strongly associated to memory. For an AI system such cleanup could be
setup daily, or simultaneously while it is awake.

George Bajszar


Sent via Deja.com http://www.deja.com/
Before you buy.






More information about the Neur-sci mailing list

Send comments to us at biosci-help [At] net.bio.net