Re: Why Human STM is Biologically Impossible

On 17 Dec 2007 00:13:06 GMT, curt@xxxxxxxx (Curt Welch) wrote:

Should we ignore the fact that there is conditioning happening and instead
look at how to implement Louis's view of STM which he says will "remember"
any 7 items including their temporal relationships?

Do you think talking about human behavior in terms of our ability to
remember a sequence of symbols is actually more useful to understanding
what we have to implement than talking about conditioning?

ahahaha... So you think that you can implement a reinforcement
mechanism for operant conditioning without first solving the STM
problem? ahahaha... You guys make me laugh. You behaviorists are out
to lunch on this one. You've been equating intelligence with
reinforcement learning for so long, you think it's the be-all of
intelligence. Look, operant conditioning is a trivial problem when
applied to a given behavioral sequence. There are several things that
must be in place in a complex intelligence before one can even begin
to think of how to implement a reinforcement mechanism.

A proper STM-episodic memory system is the sine-qua-non of operant
condtioning. Episodic memory IS the attention and behavior selection
mechanism. Unless a particular behavior is selected, no reinforcement
can take place. And before a behavior can be selected for output, it
must be brought into the selection window, i.e., it must be awakened.
And guess what? The selection window is none other than episodic
memory. That's right, the seven-item STM system is the selection
window.

One more thing. Before a behavior can be selected, it must exist to
begin with. Sure, creating behaviors is a trial and error process but
behaviors are not constructed with reward and punishment (pain and
pleasure). Reward and punishment are strictly for selection. Before
you can select a behavior you must first get rid of all the motor
conflicts in that behavior. Motor conflict resolution can be done
internally without externalizing the behavior.

Louis Savain

Rebel Science News:
http://rebelscience.blogspot.com/
.

Relevant Pages

  • Re: Is all learning due to reinforcement?
    ... >> without any reinforcement. ... If one trains a pigeon to ... It is an illustration of how STM could involve the acquisition ... certain consequences (also, perhaps, due to natural selection, though the ...
    (comp.ai.philosophy)
  • Re: The driving force of evolution
    ... idiot. ... have a phenotype, let's see if we can rationalize its descent. ... breeders who control the selection forces, but in the wild such forces ... of reinforcement that created the new behaviour, ...
    (comp.ai.philosophy)
  • Re: using Cmd-O in Firefox
    ... I use Firefox. ... Using Cmd-O opens the selection window in ... or not disappear when the file selection is made? ...
    (comp.sys.mac.apps)
  • Re: using Cmd-O in Firefox
    ... I use Firefox. ... Using Cmd-O opens the selection window in ... or not disappear when the file selection is made? ...
    (comp.sys.mac.apps)
  • Re: selection in a pic
    ... and then moving freely the selection window to some specific place in ... that pic without of course modifying the image. ... After creating the selection, choose the move tool and press the ... "selection" button under `Affects` in its tool options. ...
    (comp.graphics.apps.gimp)