Tag Archives: shaping

The 4 DogSmith Learning Games

by Angelica Steinker

 

1. Shaping

2. Prompting

3. Capturing

4. Observational Learning

 

Shaping – the process of training your dog by rewarding a very small part of the end behavior. Gradually, over time, you will require more of the dog until you have ‘shaped’ the goal behavior. For example, you want your dog to wave. You start by clicking and rewarding the dog for shifting her weight off her left paw. Next, you click and reward the dog for lifting the left paw. Next, you require that the dog lift the paw 2 inches. Finally you only click and reward if the dog has lifted her paw to eye level. At this point, you add the cue ‘wave.’

Dog howlingPrompting – the process of training your dog by using some sort of physical prop to get the goal behavior. You will be using your hand target, a stick, and other objects to help create the goal behaviors. For example, to teach the dog to spin in a circle you can first teach the dog to touch her nose to the end of a stick. You can then simply use this stick to teach her to spin by slowly moving the stick in a circular motion. The dog will want to follow the stick in order to touch her nose to it, and you will have gotten the spin that you wanted. Once you have the entire spin, add your cue. Gradually fade the stick by shortening it and then only using your hand. Eventually you can fade your hand movement and only use the verbal cue.

Capturing – is the easiest of these three methods. It only requires good observational skills and good timing. Capturing is simply clicking and rewarding your dog for a behavior that she frequently engages in. In order to capture a behavior you must be able click and reward it several times a day or ideally within the same training session. Avoid attempting to capture behaviors that only occur on an infrequent basis, the result is likely to be a very lengthy process.

Observational Learning – is the process of one being observing another and then imitating the behavior to gain access to a reward. Dogs can learn from humans via observational learning if a human touches a target, a dog that has a history of training, will likely also touch that target. Likewise dogs will mimic the same body part so if you touch the target with your hand they likely use their paw. If you touch the target with your nose they will likely touch the target with their nose too.  

Reward Delivery

Generally it is a good idea to use food for positional tricks or tricks that only require small movements such as paw and nose targeting. Tricks that require faster movement may require rewarding with a toy or a tossed piece of food. Either way both the timing of the reward, when you reward, and the delivery of the reward, where you reward, are extremely important. The reward delivery needs to support the behavior you are training. If you are teaching ‘wave’, the dog is generally asked to sit before waving. Click while the dog’s paw is in the air and then very quickly give the dog her treat while she is still sitting. Failure to do this could result in you losing both your sit cue and your wave trick. You get what you train, not what you want!

Two Types of Behaviors

Before you teach your dog a trick, decide if you want the behavior to have duration or just be a moment long. If you are training your dog to sit in a cart and then allow a person to pull that cart, you will need an on/off switch for the sit behavior in the cart. The dog sits in the cart, until you release her. Other tricks, such as left (dog spinning in circle to left) do not require an on/off switch; the dog simply does the behavior and is clicked and rewarded.

It is recommended that tricks like ‘wave’ be taught with an on/off switch. The cue wave signals to the dog to raise the paw. The cue ‘okay’ releases the dog from waving, and signals that the trick is over and she can lower the paw.

Nose and Paw Are Connected

It isn’t magic, but it can seem like it, because when you click and treat a dog for nosing a target and then you suddenly stop, the dog will usually use her paw to smack at the target. The same is true in reverse, if you repeatedly click and reward a dog for targeting with her paw and suddenly stop, the dog will usually nose the object.


Can Your Dog Fetch a Soda For You From the Refrigerator?

Written by Niki Tudge

Let us consider Sea World. I am sure many of you have watched dolphins or whales perform many creative and amazing tricks. How do the marine trainers get these behaviors and put them under stimulus control? They certainly do not use a choke chain or physically manipulate the marine mammals to jump through hoops. Dog Trainers use the same techniques to train complicated behaviors. They involve the animal in the learning process by encouraging them to voluntarily exhibit small behaviors that the trainer reinforces.Reinforcing small approximations of a desired behavior in succession to achieve a more complicated behavior is called shaping. Examples in the dog obedience world are ‘drop on recall’, ‘retrieve over a jump’, distance handling such as a ‘sit’ or ‘down’ from a distance of 30 feet. A great ‘at home’ example would be to shape your dog’s behavior to go to the refrigerator and retrieve a can of soda for you. And yes, my husband put a stop to this training session pretty quickly.

Effective behavior shaping should begin with a written plan of the behavior you desire to shape and a detailed understanding of each behavior approximation to be reinforced. Your goal should always be to remain flexible throughout the shaping session. When shaping a behavior you start by reinforcing small steps and giving immediate reinforcement for those small steps. The reinforcers are small and easily delivered so the shaping is not delayed and the dog does not get satiated too quickly. Once the dog can easily complete the first approximation of the behavior then the trainer moves on to the next approximation. We “up the ante” by expecting more from the dog to earn the reinforcement as the shaping progresses. If you put the first approximation on extinction by holding back reinforcement you will see the dog experiment with behavioral variability and they will try lots of things to keep the reinforcement coming. When you observe this you can choose the next approximation to reinforce.

Shaping is most effective when the best available approximation is reinforced. Even with a well laid out shaping plan it may be necessary for the trainer to back up or reinforce an approximation that was not on the initial plan. Lowering the reinforcement standard can sometimes accelerate learning. When the “going gets tough” use a cue and/or make smaller approximations available for reinforcement so the dog does not lose interest or shut down. Any simple mistakes you make in shaping will affect the rate of learning. If you hesitate before delivering the reinforcement, expect too much, too quickly, or inadequately reinforce the dog’s performance learning may not progress. Like all training strategies contingency and contiguity are important, as is the magnitude of the reinforcement available.

When shaping, learning is affected more by the trainer’s ability to judge and reinforce approximations than the subject’s ability to learn so if you encounter difficulties in shaping a behavior take a step back and look at what you, the trainer are doing. If you can acquire the skill of shaping then you can train a dog to do any behavior that they are physically capable of doing. Joining a well run dog training class will help you with your training mechanics, your ability to mark and reinforce the correct behavior and as such your dog’s learning will be more enjoyable and more efficient.

Niki Tudge is the founder and owner of  The Dog Behavior College and The DogSmith Franchise System. Niki is a certified NADOI Dog Trainer, a Certified Pet Dog Trainer through the Certification Council for Dog Training and a Dog Behavior Consultant with the Canine Behavior Center. Niki also holds a Diploma in Animal Behavior Technology and is a recognized Professional Dog Trainer through the Association of Animal Behavior Professionals. You can reach Niki at www.DogSmith.com or wwwDogBeahviorCollege.com