Clean Run Magazine - March 2007 - 46

 Warning : session_start : The session id contains invalid characters,
valid characters are only a-z, A-Z and 0-9 in
/mnt/data/www.nxtbook.com/fx/config_1.3/global.php on line 9 Warning :
session_start : Cannot send session cache limiter - headers already sent
output started at /mnt/data/www.nxtbook.com/fx/config_1.3/global.php:9 in
/mnt/data/www.nxtbook.com/fx/config_1.3/global.php on line 9 Warning :
Cannot modify header information - headers already sent by output started
at /mnt/data/www.nxtbook.com/fx/config_1.3/global.php:9 in
/mnt/data/www.nxtbook.com/fx/config_1.3/global.php on line 10 Clicker
Complexities Part1 By Pamela Reid, Ph.D. Nowadays, most every agility
trainer understands rudimentary clicker training. At a recent trial, I
asked a few competitors about clicker training and heard back that the
clicker is a “marker” that helps the dog recognize the correct
behavior or a “bridge” that promises a reward is forthcoming. Most
knew that the clicker should precede the reward. The majority relayed that
an effective click needs to be well timed. Everyone was pretty adamant that
each and every click should be followed by a reward of some kind. Beyond
these few details, however, people’s appreciation for clicker training
started to get a bit sketchy. Some were unsure about whether the click
should mark the successful completion of the behavior or could function as
a “keep going” signal. A few reported they use the click to signal
correct weave entries but then waffled on whether they should give a
reward if the dog subsequently made an error executing the poles. Several
trainers use the click to signal a variety of reinforcers, while others
specifically link the clicker to food rewards. So despite the popularity
of clicker training, there is still confusion about how conditioned
reinforcement works. My goals here are to clarify some common
misconceptions and delve into a few fascinating intricacies that should
prove new to many readers. Historical Perspectives Ivan Pavlov was the
first to demonstrate conditioned reinforcement in his seminal work on
respondent conditioning. Dogs learned to associate a variety of laboratory
stimuli, including the sound of a metronome and the arrival of a researcher
wearing a white lab coat, with the subsequent delivery of meat powder. The
dogs learned to salivate to these initially insignificant stimuli in
anticipation of the food. While Pavlov called these “conditional” or
“conditioned” stimuli, ensuing research confirmed that they also
acquired reinforcing properties; hence, they were also called
“conditioned reinforcers.” Early conditioning research in operant
chambers established that rewards must be delivered immediately after the
behavior occurs. Grice 1948 found that learning was impaired with delays
as short as 0.5 seconds; however, more recent experiments show
considerable learning with delays as long as 30 seconds. One classic study
examined rats learning to press a bar, with delays to reinforcement for
some rats as short as 2 seconds, for others as long as 64 seconds.
Learning dropped off fairly rapidly with increasing delays, although bar
pressing still occurred pretty 46 reliably with delays of 8 seconds
between pressing the bar and getting the food. Still, the laboratory
environment is strikingly more impoverished than the real world; a sterile
environment permits the animal to make delayed associations. In the real
world, there are typically so many things happening to interfere that it
is very difficult for the animal to associate his actions with a delayed
reward. In dramatic contrast to these studies, researchers who watched
rats running in T-mazes never witnessed the same deterioration in
performance with delays to reinforcement. Rats easily chose the correct
side of a T-maze with a 60-second delay between entering the goal box and
receiving reward. How could that be? The researchers determined that the
rats were using features in the maze, such as marks, dents, and variations
in the surface as conditioned reinforcers to bridge the delay to reward. If
measures were taken to remove unique features of the maze that the rats
could use as cues, the rats failed to learn with delays of only 0.5
seconds to reinforcement. This work revealed that animals use Clean Run |
March 07   Warning : Unknown : The session id contains invalid characters, valid
characters are only a-z, A-Z and 0-9 in Unknown on line 0 Warning :
Unknown : Failed to write session data files . Please verify that the
current setting of session.save_path is correct /var/lib/php/session in
Unknown on line 0
Clean Run Magazine - March 2007

Table of Contents for the Digital Edition of Clean Run Magazine - March 2007
