| View previous topic :: View next topic |
| Author |
Message |
baylor
Guest
|
 Posted: Sat Aug 16, 2003 7:54 am Post subject: How does adaptation work? |
 |
|
Anyone know how people quickly adapt in learning?
In AI, you have reinforcement learning where you find a value by
basically averaging the results of all iterations, all results weighed
equally. So if you touch 10 burners and don>t burn your hand and then
touch one and get a severe burn, the next time you see a burner you>ll
asume it>s safe. And that ain>t how humans work
For some reason i thought this was governed by the sum of experience
with extra weight given to recency and impact, but i can no longer
figure out where i came up with that idea and i can>t figure out the
proper search terms to use to make Google cough up some useful
information (i keep getting WM recency and primacy results which ain>t
the same thing)
Any one got any advice?
-baylor |
|
| |
|
 Back to top |
Randolph M. Jones
Guest
|
| Posted: Mon Aug 18, 2003 8:22 pm Post subject: Re: How does adaptation work? |
|
|
It appears to me that finding out something about it is exactly what he
is trying to do.
Glen M. Sizemore wrote:
[quote]You are asking questions that have been actively, and directly, investigated
for almost 70 years. If you want to simulate it, you should know something
about it.
http://www.envmed.rochester.edu/wwwrap/behavior/jeab/jeabhome.htm
"baylor" <baylor@no_spam.ihatebaylor.com> wrote in message
news:c07rjvoo2fl97oo8pnpk9850gmhqo12h4n@4ax.com...
Anyone know how people quickly adapt in learning?
In AI, you have reinforcement learning where you find a value by
basically averaging the results of all iterations, all results weighed
equally. So if you touch 10 burners and don>t burn your hand and then
touch one and get a severe burn, the next time you see a burner you>ll
asume it>s safe. And that ain>t how humans work
For some reason i thought this was governed by the sum of experience
with extra weight given to recency and impact, but i can no longer
figure out where i came up with that idea and i can>t figure out the
proper search terms to use to make Google cough up some useful
information (i keep getting WM recency and primacy results which ain>t
the same thing)
Any one got any advice?
-baylor
[/quote] |
|
| |
|
| Back to top |
Glen M. Sizemore
Guest
|
| Posted: Tue Aug 19, 2003 1:18 am Post subject: Re: How does adaptation work? |
|
|
Just thought someone might want to find out how real animals and real
reinforcement works.
"Randolph M. Jones" <rjones@colby.edu> wrote in message
news:3F40EF23.4030402@colby.edu...
[quote]It appears to me that finding out something about it is exactly what he
is trying to do.
Glen M. Sizemore wrote:
You are asking questions that have been actively, and directly,
investigated
for almost 70 years. If you want to simulate it, you should know
something
about it.
http://www.envmed.rochester.edu/wwwrap/behavior/jeab/jeabhome.htm
"baylor" <baylor@no_spam.ihatebaylor.com> wrote in message
news:c07rjvoo2fl97oo8pnpk9850gmhqo12h4n@4ax.com...
Anyone know how people quickly adapt in learning?
In AI, you have reinforcement learning where you find a value by
basically averaging the results of all iterations, all results weighed
equally. So if you touch 10 burners and don>t burn your hand and then
touch one and get a severe burn, the next time you see a burner you>ll
asume it>s safe. And that ain>t how humans work
For some reason i thought this was governed by the sum of experience
with extra weight given to recency and impact, but i can no longer
figure out where i came up with that idea and i can>t figure out the
proper search terms to use to make Google cough up some useful
information (i keep getting WM recency and primacy results which ain>t
the same thing)
Any one got any advice?
-baylor
[/quote] |
|
| |
|
| Back to top |
Glen M. Sizemore
Guest
|
| Posted: Tue Aug 19, 2003 3:33 am Post subject: Re: How does adaptation work? |
|
|
Just thought that someone might wonder about real animals and
real reinforcement.
"Randolph M. Jones" <rjones@colby.edu> wrote in message news:<3F40EF23.4030402@colby.edu>...
[quote]It appears to me that finding out something about it is exactly what he
is trying to do.
Glen M. Sizemore wrote:
You are asking questions that have been actively, and directly, investigated
for almost 70 years. If you want to simulate it, you should know something
about it.
http://www.envmed.rochester.edu/wwwrap/behavior/jeab/jeabhome.htm
"baylor" <baylor@no_spam.ihatebaylor.com> wrote in message
news:c07rjvoo2fl97oo8pnpk9850gmhqo12h4n@4ax.com...
Anyone know how people quickly adapt in learning?
In AI, you have reinforcement learning where you find a value by
basically averaging the results of all iterations, all results weighed
equally. So if you touch 10 burners and don>t burn your hand and then
touch one and get a severe burn, the next time you see a burner you>ll
asume it>s safe. And that ain>t how humans work
For some reason i thought this was governed by the sum of experience
with extra weight given to recency and impact, but i can no longer
figure out where i came up with that idea and i can>t figure out the
proper search terms to use to make Google cough up some useful
information (i keep getting WM recency and primacy results which ain>t
the same thing)
Any one got any advice?
-baylor
[/quote] |
|
| |
|
| Back to top |
|
