|
Resources > Training
Operant Conditioning (Under construction)
How do we get our canines to do the complex behaviors of SAR work?
Finding missing people sounds complicated. Is there a preferred breed or intense training program we put our
canines through? No. All of our dogs are trained by their owners through operant
conditioning utilizing positive reinforcement. Huh, what does that mean? Can I
do this with my dog without getting confused? ABSOLUTELY! To learn more, keep
reading but keep the following in mind: Every interaction you have with your
animal has an impact on its behavior; training does not stop after you put away
the clicker; with time and patience you can train anything!
Classical conditioning overview.
To better understand current
positive training methods, you should be aware of classical conditioning. This
is an involuntary or automatic response to a stimulus. Ivan Pavlov (1849-1946)
discovered conditioned reflexes and stated that the organism has little control
over the response and that the environment elicits the behavior from the
organism. The relationship between behavior and environment is
stimulus-response. Perhaps you are familiar with Pavlov's dog experiment? In his
experiment, Pavlov rang a bell before feeding a dog. Over time, he noticed the
dog salivated after hearing the bell ring and eventually he could ring the bell
without feeding the dog and the dog would still salivate. Why? The dog
involuntarily learned that the bell meant food and anticipated its arrival by
salivating. This is also known as respondent conditioning.
Operant
conditioning overview.
Modern conditioning principles
are similar to classical conditioning but voluntary behavior is emitted by the
organism, which is controlled by its consequences. This relationship was defined
by BF Skinner in his book, The Behavior of Organisms, (1938) as a type of
learning in which the probability of a behavior recurring is increased or
decreased by the consequences that follow it. The animal/person being
conditioned is the operator AND chooses to engage in the activity.
Believe it or not, this is an easy way to teach and an easy way to learn. Simply
stated, operant conditioning is modifying behavior through consequences. Much of
our learning is operantly conditioned- we make choices about our actions based
on our past experiences. Our goal is to teach animals to respond consciously
(not react as with classical conditioning) to each of our interactions with
them. Once you understand the basic theory behind operant conditioning, the
application (or practicing) of these techniques does not seem so difficult.
Ten laws of operant conditioning.
Karen Pryor is a leading animal behaviorist and author of Don't Shoot the Dog! (A must read for the
serious trainer!) In her book, she defines ten important rules for operant conditioning. They are as follows:
- Raise criteria in small increments.
- Train one criterion at a time.
- Vary reinforcement before moving to the next approximation.
- Relax old criteria when introducing new criteria.
- Plan ahead.
- Don't change trainers mid stream.
- If a plan does not work, change the plan.
- Don't stop a session gratuitously.
- Regress when behavior deteriorates.
- End on a positive note.
What is a bridging stimulus?
A stimulus that pinpoints in time the precise moment of a desired behavior and bridges the gap in time
between that point and when the organism may receive further reward. This is the MOST important piece in training.
A bridging stimulus, or bridge, is also a conditioned/secondary reinforcer or a
marker that means, "yes", "good", "correct", "food is on the way". Think of it
as a universal language, communicating to your animal that it has done something
in/correctly. If s/he hears the bridge then s/he performed correctly; is s/he
does not hear the bridge then s/he has not performed correctly.
The bridge helps in the timing of reinforcement and can be one of many secondary
reinforcers for your animal, this one just has a more specific meaning. Each
bridge must be precise and have consistent clarity. Remember, this is your
communication to your canine!
There are 3 types of bridges- verbal, whistle, and clicker. The clicker is the
more popular; they are sold in almost every pet store and all sound the same.
Once you have one, how do you begin using it? Start by associating the bridge
with a primary reinforcer, like food. You can place a piece of your canine's
food on the floor and bridge each time s/he eats it. Your canine will quickly
learn that the clicker equals food. You can practice your bridging timing by
catching keys in your hand and bridging when they immediately touch your hand.
What is stimulus control?
A discriminative stimulus (indicated as SD but also known as a cue or command) is a stimulus that precedes
a behavior, signaling that a specific response will be reinforced if emitted
correctly. The result is that the stimulus will consistently elicit only that
particular response. There are a few options to choose from. You can use a
visual cue (use your hands to tell your dog to sit), an auditory cue (verbally
tell your dog to sit),or a tactile cue (touch your dog with a specific pat on a
specific body part to tell your dog to sit). Whichever type of SD you choose to
use, make sure you are consistent with it. A good rule of thumb is to use one
cue for each behavior. Don't put your animal in a position where it has to guess
what you are asking for; remember s/he doesn't understand English!
When a trained behavior occurs consistently in response to an appropriate SD it
is considered to be under stimulus control. The behavior should meet four conditions:
It is immediately offered following the SD It is offered only when preceded by the correct SD It is not offered in the presence of another SD It does not occur in the absence of the SD.
Are there different types of reward systems?
Consequences are highly effective tools in the learning process. There are two types of consequences-
reinforcement and punishment. Reinforcement is the preferred consequence in
training. Punishment should be reserved by those who fully grasp its detrimental effects.
Reinforcement is any circumstance/event that INCREASES the frequency of that
behavior occurring again. This is the basis of all conditioning. Timing is key;
if reinforcement is delivered too early or late the animal might be reinforced
for a different behavior or forget what behavior it was reinforced for.
Punishment is the removal of the opportunity to obtain reinforcement, which
DECREASES the frequency of that behavior occurring again. Punishment can also be
positive (adds something to the animal's environment) or negative (removes
something from the animal's environment). This consequence should only be used
by advanced trainers who fully comprehend its detrimental effects.
How is reinforcement used?
There are three types of reinforcement: Positive (adds something to the animal's environment), Negative
(removes something from the animal's environment), and Differential. Positive
reinforcement is the most common tool in teaching your animal s/he has done
something correctly. Negative reinforcement usually results in the animal
looking to escape or avoid the situation. Differential reinforcement is when one
behavior is reinforced and others are not. This is most commonly used for
incompatible behaviors- your dog cannot physically sit down AND jump on you at
the same time. If you want your dog to sit when you walked in the door instead
of jump up, you would differentially reinforce the sit behavior.
Reinforcers are anything that occurs in conjunction with a behavior that tends
to increase the likelihood that the behavior will occur again. A primary
reinforcer is something that the animal finds inherently reinforcing, i.e.
something that satisfies a basic biological need- food. A secondary reinforcer
is something that has acquired reinforcing value through its association with a
primary reinforcer- a toy, tactile, praise.
There are different ways of applying reinforcers. These are known as schedules
of reinforcement, of which there are two types. A continuous reinforcement
schedule is given after each correct response. An intermittent reinforcement
schedule is given periodically after a certain number of responses.
What do you do when your animal performs an incorrect behavior?
You should ignore the behavior
and not let you emotions get it in the way. A least reinforcing stimulus (LRS)
is a calm, neutral response of 3-5seconds that occurs at the time of the bridge
(had there been one). This lets the animal know s/he performed an undesired
behavior but that it is OK to make a mistake. Both animal and trainer should
remain calm. This method can be used numerously and safely as a learning tool.
If you act on the incorrect behavior, chances are you are using negative
reinforcement and/or punishment. A time out (TO) is the removal of opportunity
for positive reinforcement. This should only be used once or twice to decrease
an animal's angry or frustrated state of mind. Extinction is a method of
eliminating a behavior by no longer reinforcing it. This tool should also be
reserved for the more advanced trainers as frustration and aggression are common
"side effects".
Is there an easy way to get rid of undesirable behavior?
Karen Pryor lists some methods to get rid of undesirable behavior in her book, Don't Shoot The Dog! How many
have you thought about or tried?
Shoot the animal Punishment Negative reinforcement Extinction Train an incompatible behavior Put the behavior on a cue Shape the absence of unwanted behaviors Change the motivation
What is superstitious behavior?
A behavior that the animal offers during the training of another behavior but is unrelated to the behavior
being trained. Because the unrelated behavior is inadvertently reinforced with
the desired behavior, the animal perceives it as a necessary component of the
behavior being trained and therefore necessary to receive reinforcement. As the
trainer, you must be constantly aware of this possibility, as it can lead to
problems later in training.
Why do superstitious behaviors occur? There are several possible reasons. Poor
bridging technique, such as poor timing or too long of a bridge thereby
capturing more than one behavior. Poor observations. Sloppy cues which the
animal could learn over the original cue.
How do I begin now that I've learned some basics?
Before you dive into this exciting world, take a few minutes to asses yourself and your animal.
Take an animal history. Do you know what type of dog you have? This may seem
like a silly question but for those of you with mixed breeds, this will help
figure out several behavior questions. Do you know its background? Was it
previously abused? Was it hand raised or parent raised? Animal trust goes both ways. Does your animal trust you? How well do you
trust your animal? Spend some quality time bonding so you both learn something
about each other. Learn those operant conditioning techniques. The tip of the iceberg was
covered here, so go soak up more knowledge as you progress with your training. Learn the bridge. Remember, this is how you will communicate with your animal and is the most important piece in training! Learn what a great tool positive reinforcement is and become proficient in using it. Remember the behavior you train are extensions of natural behaviors, even the crazy, fun stuff, so be creative. Your SD must be consistent and different from other cues you are using. Don't start training without a plan and use successive approximations (small steps) to train. Too many big steps forward can get confusing quickly! Be creative- make sessions fun but know when to stop and ALWAYS end on a positive note.
This site was last updated on
04/23/2007
|
