Читать книгу Abc's Of Positive Training - Miriam Fields-Babineau - Страница 9

The first people to use operant conditioning outside of the laboratory environment were Keller and Marian Breland. They were students of B. F. Skinner during the 1940s. During this time, they used operant conditioning on dogs. Keller Breland was the first to train marine mammals during the 1950s and, soon after, marine park shows were born, such as those at Marineland and Sea World. Operant conditioning is used to this day to deliver to audiences the spectacular dolphin and whale performances seen at amusement parks.

Animal trainers have taken operant conditioning to a more advanced level, pairing it with classical conditioning. Think Pavlov’s dog: a bell is rung, which signals food, thus causing the dog to salivate. The dog has learned that the sound brings a food reward. The release of the food reward is preceded by a bridging signal which can be either a light, a buzzer or a clicking sound. The dog learns that a correct response brings the bridge, which signals that the food is sure to come. This allows a trainer to shape multiple behaviors quickly. Here are a few examples of positive reinforcement:

• A dog gets a biscuit for going potty outside;

• You get paid for doing a job;

• A child gets to go to the amusement park for earning good grades;

• A dog receives petting when he jumps up on a person;

• You receive a raise for a job well done.

Marine-mammal trainers use a whistle to bridge the correct behavior.

Marine mammals respond well to visual cues.

When a dog jumps up, it’s best to not let him get his paws on you at all, as this contact serves as a reward for him.

These examples are all primary reinforcers. A primary reinforcer is a reward that the receiver doesn’t have to learn to like. There are also secondary reinforcers. These are actions that the receiver does have to learn to like. Examples are:

• Releasing a tight leash when the dog stops pulling;

• A child is no longer yelled at when his room has been cleaned;

• A cow is no longer shocked (with an electric cattle prod) as long as it keeps moving forward.

Both primary and secondary reinforcers are utilized in animal training as well as in our own everyday lives.

A term with which you should be familiar is “schedules of reinforcement.” There are several types of schedules of reinforcement:

• A fixed interval: a reward will be given after a fixed amount of time; for example, every two minutes or every ten minutes;

• A variable interval: a reward will be given but the receiver has no control over when;

• A fixed ratio: a reward will be given after a specific number of correct responses;

• A variable ratio: a reward is given when the greater number of correct responses occurs within a specific set of stimuli;

• A random interval: there is no correlation between the number of correct responses and the receipt of the reward.

Another term with which you should familiarize yourself is “extinction.” This is when a behavior has not been reinforced, thereby fading from the repertoire. This is a good means of getting rid of a behavior without having to use punishment, which in itself might encourage a behavior if it is the only means of obtaining attention. An example of this is yelling at a dog for barking, which encourages the behavior by the owner’s taking part in the “game.” Another example is pushing at the dog for jumping up. The dog is being touched, thereby receiving reinforcement. To extinguish these behaviors (barking and jumping up), they would need to be ignored. Granted, it’s difficult to ignore either of them, but the dog’s learning that the behaviors no longer bring him pleasure of any kind will eventually extinguish the behaviors.

Skinner defined four possible ways to modify a behavior: positive reinforcement, negative reinforcement, positive punishment and negative punishment. Reinforcement is the use or removal of a stimulus to increase the occurrence of a behavior. Punishment is the use or removal of a stimulus to decrease the occurrence of a behavior.

Your reaction to a behavior will tell the dog whether or not to continue the behavior. Let’s say that your dog is digging in the trash. To him, the reward is getting something to eat, so he will repeat the behavior whenever he gets the chance. However, you can do one of two things to get rid of the behavior: you can either remove the trash can and place it somewhere that Buster can’t access it, or you can punish him for the behavior by introducing something aversive, such as a “scat mat.” A scat mat has electrical stimulation that will cause the dog’s toes to tingle when he steps on it; hence, he’ll quickly learn to avoid the scat mat and the trash can.

Removing the trash can is considered negative punishment. This is when something that had been rewarding to the dog is taken away, which will reduce the occurrence of the behavior. The scat mat is considered positive punishment. This is when something aversive is introduced during the behavior that creates enough distress to stop the dog from performing the behavior.

A secondary reinforcer would be the release of the pressure of the leash when the dog stops pulling.

The positive-reinforcement way to stop a dog from digging in the trash is to divert his attention, have him sit and then click/reward.

There are other ways to handle the trash-digging pup, and that is through positive reinforcement or negative reinforcement. If you were to divert the dog’s attention from the trash can, have him sit instead and then give him a treat, you’d be offering positive reinforcement. If you were to have a leash on your dog and gave it a yank when he went near the trash can, this would be considered negative reinforcement. There are two types of negative reinforcement: avoidance and escape. In avoidance, the dog would steer clear of the trash can when walking in the same room due to the threat of being yanked. In escape, he would run from the room when he even sees the trash can.

Which do you think would be most effective in teaching Buster to stay out of the trash? Actually, it’s a combination. Using just one method might cause the dog to get the wrong idea or to become overly frightened of you. The type of conditioner you use depends largely on the dog’s personality, the situation in which the behavior is performed and what your ultimate goals are for your pet.

There are several things to consider when using operant conditioning to train your pet. First of all, you have to make certain that you reward the behavior that you want. For example, if your dog is growling at someone due to fear, the last thing you should do is pick him up and speak in a soothing tone of voice. This reinforces the growling behavior by positive reinforcement and encourages the dog to continue. Instead, negative punishment would be more effective. Using this method, he does not receive attention for his growling; instead, he is not allowed to be with you. Positive punishment might also be more effective. Let’s say that when he growled, you squirted him with water. Buster learns that growling earns him a squirt in the face instead of the rewards of being picked up and spoken to in pleasant tones. This will extinguish the behavior.

Another thing you must be aware of when using positive reinforcement is timing. Timing is everything. If you reward your pet at the wrong moment, you will be reinforcing the wrong behavior. For example, you ask your dog to sit. He sits, then gets up. You have not yet taught him to stay and you haven’t rewarded him for doing what you asked. He’s not sure what to do to earn his reward. If your timing were correct, he would’ve been bridged (signaled at the moment he sat) and given the reward before he got up. This way, he would correlate the sit action with his reward and the behavior would be repeated. Otherwise, he would correlate the action of getting up with his reward and you’d have a dog who has learned to not sit. The bridge and following reward need to be done at the very second that the dog does what you’ve asked of him.

It’s the rare dog who is not motivated by food, although it does happen. Most dogs have a good food drive and will perform for a tasty reward.

Another factor to consider is the value of the reward. Dogs differ in the values they have regarding what motivates them. Some are happy with a touch or kind words, while others need pieces of hot dog or liver. You’ll need to discover what motivates your dog prior to beginning any training. If your dog likes a number of things, you can vary his rewards according to how well he performs. Let’s say you are teaching your dog to sit in heel position. When he sits at your side, but not in heel position, he receives verbal praise but no treat. When he sits a little closer to your side, facing forward, he receives a piece of biscuit. As he learns that better things come to him the closer he sits, he’ll sit straighter in proper position. As he achieves his goal, his reward is a piece of freeze-dried liver or hot dog. Buster has learned how to achieve without your having to force him into it. It would be similar to your getting a better raise at work due to putting in more hours than normal versus a regular raise for merely consistent performance.

The Round Robin game is a great way to involve all family members in the training process, as the dog goes from one person to the next and performs what is asked of him.

A problem with training in this manner is that the animal can learn to associate the reward with only the trainer. He will only listen to or work for this one person. In order for the animal to respond to the entire family, everyone will need to work with Buster. This way, he learns to listen to everyone because anybody in the family has the possibility of being a reward-dispenser. However, when teaching something new and complex, it’s a good idea for just one person to be the key trainer so as not to confuse Buster. Once the behavior is learned, others can join in.

Teaching targeting with a target stick.

This hopeful beggar seems to ask, “Do I get a treat now?”

If your pet is being rewarded with food, he can become sated. Once sated, he’ll no longer be motivated to perform. In order to avoid this, very small pieces of treats should be used, or he can learn to perform for his regular meals. This way, you’ll be able to maintain his proper eating habits and weight while he “works for a living” just as he would in the wild. This would so closely resemble his natural instincts that the dog would feel very fulfilled with his learning process.

Some dogs tire of the same type of food reward, so it would be a good idea to vary the types used. If using freeze-dried liver, try different flavors. If using hot dogs, try different brands or substitute some bacon once in a while. Popcorn is usually a favorite, as are small bits of pizza. As with any treats, dispense in moderation and make sure you choose something that doesn’t irritate your dog’s digestive system. Inflamed bowels can cause extinction in the behaviors you’ve achieved, because the dog might learn that the more he performs, the worse he feels.

In order to control when your dog performs, utilize only positive reinforcement when teaching a new behavior during a training session, otherwise your dog will constantly be bugging you by throwing out behaviors until he receives his reward. At first, the behaviors will be those previously taught, but when he sees that he isn’t rewarded for them, the behaviors will turn to negative ones that will ensure him of attention of some kind.

Puppies have short attention spans and are easily distracted. You will need a reward with high value to the pup to keep his attention focused on the lesson; even at that, you should only work with pups in several-minute sessions.

In everyday life with your dog, there are chances of creating what is termed “secondary negative punishers.” We tend to do this inadvertently and need to strive to either avoid these situations or turn them into something positive. An example of a secondary negative punisher is calling your dog to come and thus taking him away from something he was enjoying, such as leaving a doggy playgroup or ending a fun game of fetch. While there’s really no means of avoiding these situations, you will need to identify them and try to turn them into something positive. So the next time you call your dog to come inside from digging in the yard, think of offering him something that might be more enjoyable, such as eating his meal or doing some tricks for food rewards. Even a tummy rub would be great!

Positive punishment alone can also be abusive if not used correctly. Pairing positive punishment with a secondary positive punisher would be more humane and teaches the dog to correct himself upon hearing the trigger word, such as the word “No” along with a punishment. An example is to say “No” as the dog gets a squirt of water in the face. The dog learns that “No” will coincide with the punishment, thus reducing his bad behavior as soon as he only hears the word. This form of punishment is a very humane means of correcting your dog, although there are some schools of thought in which they strive to not punish at all, in which case the undesired behavior may or may not extinguish.

To a retriever, a game of fetch is very rewarding.

Giving the dog a chew toy to play with is a positive way to end a training session.

Some dogs learn that while a specific behavior doesn’t receive a reward from their human companions, the behavior itself is rewarding. Digging is self-rewarding, as is obtaining scraps from the garbage can. Barking is a fun means of releasing energy. Chewing is a great way to relieve anxiety.

Punishments can also be associated with a specific person. In order for them to work, they must be used by all family members. If the dog has a tendency to do something when nobody is around, he should be contained in an area where he cannot perform the behavior when unwatched. This way, he is conditioned to not behave in such a fashion at any time that he has access to the means of deploying the behavior. An example would be a dog getting on the furniture. Whenever anyone is nearby to see the dog get on the furniture, the dog is corrected for doing so through the use of positive punishment, such as a squirt of water or tug on the leash, along with a secondary punishment, such as the word “No” or “Off.” Eventually he’ll learn to either avoid the furniture or to get off when the secondary punishment is used. However, when no one is around to give the punishment, he gets on the furniture. To condition him out of this behavior, he should not have access to furniture when no one is around to watch him. Eventually the habit will become extinct.

Digging is an example of a self-rewarding behavior.

There is also another possible outcome to an animal’s behavior and that is no response. There is no reward marker, no “Good Dog,” no secondary positive punishment (“No”) and no correction. This is considered the “keep going signal,” also called the “no reward marker.” For dogs that are well into how to throw out (keep performing) behaviors in order to receive their reward, this lack of signal will cause them to continue throwing out behaviors until they get something right. A special word can be formulated to coincide with this lack of signal. Some trainers use “Nope” or “Wrong” or “Try Again.” This will cause the dog to try again. However, this will only work with dogs that have been trained via positive reinforcement and have an attention span that will prevent their extinction of performing altogether.

Animals can suffer both physical and mental damage from incorrectly applied punishments and reinforcers. Using these techniques correctly will prevent harassing or abusing your pet. Be certain that you are clear on the procedures prior to beginning the training process. Try them out on a family member or friend. The children’s game “Stop and Go” is a great way to practice, as is the game “Hot or Cold.” Things to think about when applying the techniques are: Does it work? If not, why? Was the timing off? Dogs are very forgiving, but once a behavior is learned, it is far more difficult to extinguish it than it is to teach it correctly in the first place.

If nobody is around to catch him, a furniture-hopper will certainly settle in for some couch time.