Reinforcement plays an important role in increasing the frequency of behavioral responses. But in order to use it for behavioral training, the target behavior must appear at least a few times during the training process.
On the contrary, when the dog fails to achieve the target behavior or cannot directly appear a certain behavior, the frequency of the behavior cannot be increased or reinforced.
Therefore, in this case, it is necessary to operate the conditional reinforcement principle to make the target behavior or a certain behavior appear and reinforce, that is, behavior shaping.
What is behavior shaping?
Behavior shaping evolved from Dr. Skinner’s theory of operant conditioning,
a means used to shape the dog’s inability to directly perform the target behavior, so that the The dog’s target behavior is approached, and eventually it is a step-by-step reinforcement process of this target behavior.
Behavior shaping, also known as continuous approximation method, is the approach and continuous reinforcement of a series of continuous actions to the target behavior, and finally makes the dog make the target behavior and stabilize.
Behavioural shaping is widely used in dog training
, such as
Frisbee dog training process:
Make the dog interested in toys (balls, knots)→Learn to chase the toy→Learn to pick up the toy in exchange→Train the dog to play Frisbee→Close up with the Frisbee → Take the Frisbee at a distance.
In this series of processes, a series of actions are formed.
At the same time, the dog trainer should feel happy and give encouragement to the progress of each stage in the process of the dog picking up the Frisbee. This process is reinforcement, which is a step-by-step, step-by-step process.
Characteristics of behavioral shaping
Phases of association
Behavior shaping is generally used for more complex behaviors. A complex behavior is difficult to reach the target behavior with only one step.
Therefore, in the process of behavior shaping, the final target behavior is often decomposed into interrelated staged target behaviors.
For example, the process of a dog learning to play dead can be refined into the following staged target behaviors:
The dog came to me → sat down beside me → lay down beside me → lay down beside me → lay down completely beside me → lay down beside me and held for a while → the dog still lay down when the person left keep still.
In the staged target behavior, the first stage target behavior is also the initial behavior, which is the simplest behavior training stage, and is the behavior that the dog can easily obtain.
The final goal behavior is that the person leaves the dog and still lies still and remains motionless.
So the goal behavior in the last stage is also called the end goal behavior, which is the final goal of shaping the result. target behavior.
Graduality
The target behaviors at each stage of the behavioral shaping process are interconnected , is a kind of continuous approaching behavior, only on the basis of the completion of the previous behavior, can the next target behavior be entered, thus forming a series of progressive processes.
Such as training a dog to retrieve a ball:
Take the dog to a quiet room and place a ball somewhere in the room.
The steps in this training shaping process are:
The dog looks at the ball direction → move closer to the ball → touch the ball → touch the ball with the mouth → pick the ball in the mouth → keep the ball in the mouth → walk back with the ball in hand → spit the ball into the hand of the person.
The dog can only approach the ball after looking in the direction of the ball,
It is possible to touch the ball after it has the behavior of approaching the ball, and so on. , and finally gradually reach the end target behavior.
The process of behavior shaping is actually a process that starts from the simplest and most easily acquired actions by dogs and moves step by step toward the final target behavior.
Use both reinforcement and extinction strategies
In the process of behavior shaping, The approaching behavior of each stage is actually a new behavior.
Every time the dog reaches a predetermined staged goal, the trainer must give it reinforcement.
When the reinforcement reaches a certain level, the trainer will stop the reinforcement, which often leads to the disappearance of the behavior. In fact, this process is the process of behavioral regression.
Of course, the main purpose of stopping reinforcement is to move on to the next stage of the target behavior.
For example, when the dog touches the Frisbee for the first time, the owner will hug the dog very happily, tug-of-war with him and Giving food,
this is actually a kind of reinforcement for the dog, and the dog that receives the reinforcement message will continue to exhibit similar behavior.
The first few times, the dog handler will continue to reinforce it with a timely response.
When the dog trainer feels that it is no longer necessary to strengthen the simple touch of the Frisbee, if the dog continues to appear and just touch the Frisbee, the dog handler should not respond,
Because if it continues to be strengthened, the dog will only make the simple behavior of touching the Frisbee, and will no longer take the Frisbee or return to the owner with the Frisbee.
Similarly, when the dog makes a few simple touches of the Frisbee and gets a response from the untrained dog, the dog loses interest in continuing the touching behavior.
Sometimes dog trainers use some method to teach dogs to pick up the fly.
For example, imitation learning, in front of the learning dog, reinforces the behavior of other dogs with flying discs. Some dogs will try to learn how to pick up the flywheel from other dogs in the hope of getting a positive response from the trainer.
When the dog is holding the flywheel, the dog trainer should give a positive response in time. At this point, the process is a reinforcement of the dog’s new behavior of holding a Frisbee.
