equine clicker training

using precision and positive reinforcement to teach horses and people

ClickerExpo 2016: Kay Laurence, “Hocus Focus: Helping your dog determine what’s important.”

distractions1These notes are from a lab that was taught by Kay Laurence. I took notes on each dog and then looked for overall trends or comments that would be useful and organized them into this report.

This means my report is a little different than some of my past reports because it is more of a summary of what she shared, and not a point by point account of the material. It also means that some of the content is based on my understanding of what she was doing. If you have questions, let me know and I can ask Kay for a clarification.

 
 

There were four main things that she covered in this lab. They were:

  • Distractions vs. relevant and irrelevant details. Does it matter what we call them?
  • Who determines relevant vs. irrelevant?
  • How do we teach our dogs to focus?
  • How do we teach our dogs the difference between relevant and irrelevant details?

Distractions vs. relevant and irrelevant details. Does it matter what we call them?

Kay started by telling everyone present that they were not allowed to use the word “distraction.” The word distraction comes from traditional training and has an implied meaning that the trainer must be more important than other things in the environment. Instead she prefers to talk about relevant and irrelevant details. Relevant details are the things that we want the dog to notice. They are useful information. Irrelevant details are those details that we want the dog to recognize as not providing useful information about what to do.

Here are some examples of details that we usually want the dog to consider relevant:

  • Our cues
  • Context cues that provide useful information about how to do a behavior
  • the equipment we are using (target stick, platform, etc…)
  • The click
  • The placement of the food after the click

Here are some examples of details that we usually want the dog to consider irrelevant:

  • Our posture – it’s important to be able to cue behaviors while standing in different positions (arm and leg placement and/or movement)
  • The location of the food reserve
  • our clothing, appearance, etc…
  • presence of other people, dogs, animals, etc… in the training area
  • many stimuli that dogs normally orient to (smells, trash, etc…)

Note: I made these lists based on what we did in the lab. They did not come from Kay and are not intended to be complete, just to get you thinking.

Who determines relevant or irrelevant?

We can make decisions ahead of time about whether we want something to be relevant or not, both for a particular behavior and in a particular training session. It’s important to make a list of possible significant stimuli and ask yourself what you want the dog to do if she notices them. If there are possible stimuli that we want to make sure are irrelevant, then we need to teach that distinction as we train the behavior. To do this, we set up conditions where we introduce changes and give the dog time to learn which ones are relevant and which ones are not. Kay had some exercises that showed us to do this. I’ve described them later in this report.

So it is important to plan ahead and teach your dog about relevant/irrelevant details as part of your training. On the other hand, while we are training, unexpected things might happen and we need to recognize that in those circumstances, the dog needs to be allowed to assess the change and decide if the new stimulus is relevant or not. Some things are more relevant than other things and we have to wait for the dog to decide if something has gone from relevant to irrelevant. Do we wait for the relevance to be assessed? Yes.

One example she used here was what would you do if someone walked into the room carrying a gun? Most of us would stop what we were doing and assess the situation. If we determined that it was a Marshall or a policeman who was visiting, then we would probably go back to what we were doing. The gun went from relevant to irrelevant after our assessment.

What about if you were in the room and you couldn’t see the man with the gun very clearly, but were told he was ok. Would it matter who told you? If you don’t know someone’s assessment skills, should you trust them? I think most of us would be tempted to take a quick look ourselves, just to be sure.

It seems unfair to expect dogs to accept our assessment unless we have proven to them that we can be trusted. But we can build that trust over time.  We do that by teaching them that we can give them useful information about whether something is relevant or irrelevant.   One part of the teaching process for this involves asking for already trained behaviors under varying conditions.  Another part of it comes from introducing irrelevant details so they get to practice assessing changes in the environment.

For example, you may be out walking your dog and he sees vomit on the pavement. Most dogs are going to want to check it out. If this happens to Kay when she is walking her dog, she cues the dog to “walk on.” This is a behavior the dog knows well and is incompatible with sniffing vomit. This is preferable to saying “Leave it,” which like saying “no,” because it doesn’t tell the dog what to do. Leave it and do what? Better to just say “walk on.” Leave it is another example of a word/behavior that we have inherited from traditional dog training, but that is not the clearest way to communicate with our dogs.

Let’s leave irrelevant vs. relevant for a moment and think about the underlying skill which is learning to focus.

How do we teach focus? Using a Target stick cup to teach focus:

Focus is a learned skill. It requires a lot of mental effort so it’s important to remember that focus disappears when you get mental fatigue. Some jobs require a tight focus where you are constantly filtering out irrelevant details (working in a busy office, air traffic controller). Other jobs or activities may allow for some loss of focus without it causing significant problems. It’s important to teach your dog to focus, but you need to be aware that it takes energy and you may need to alternate or vary between tasks that require a tight focus and others that do not.

One of the ways that Kay teaches a dog to focus is by taking advantage of the dog’s natural ability to follow a scent trail. She teaches this using her own style of targeting stick, which is a wooden stick (dowel) with a small measuring cup attached to one end. She likes using the metal measuring cups with slightly squarish bottoms as that helps to keep the food in the cup during movement. The cup can be attached by tape or heat shrink tubing.

The dog is taught to follow the movement of the cup. When I first saw this a few years ago, she called it a “scent target” because the dog learns to follow the scent of the food, which is a very natural behavior for a dog. It could be considered a form of luring, but since dogs were designed to smell, she doesn’t feel it’s a problem. And she finds it is very easy to use the target cup stick to prompt natural behavior which you can reinforce and put on cue.

Here are some additional points/tips on using a target stick cup:

  • Practice with the target stick cup first (without the dog). Practice moving it without the food falling out and then practice deliberately placing the food where you want it. To feed, you lower the cup and tip it out so the food is placed on or drops a short distance to the ground.
  • Slightly wet food works well as it is less likely to fall out.
  • You need to be able to load the cup while the dog is not looking. Load it while the dog is eating.
  • Teach the dog to follow the cup in movement and while standing still.
  • Using the target stick cup maintains focus better than a regular target stick because the dog is not watching your food hand or checking in with you after getting the treat. You load the cup and put it right in front of the dog so she can start again.
  • You can elicit natural behavior by how you move the cup. For example, if you lift the cup up, the dog will stop and stand because the cup movement is similar to a bird going up and triggers standing still.
  • Teaching a dog to circle with a target stick cup is easy because the dog can follow the scent trail.
  • Do short sessions, about 10-12 treats and then take a break

She demonstrated how to introduce and start working with the target stick cup to one dog. First she got him to look at it, clicked, lowered the cup and placed the food on the floor. Then she got him following it a few steps before she clicked and reinforced him. She mixed up asking for a few different behaviors so she might click a few times for following the cup and then clicking for orienting to the cup while standing or using the cup to ask for stop. She played with movement of the cup to ask for turns and stops. Her rate of reinforcement was high enough that the dog became very interested in following the cup.

She introduced the target stick cup early in the lab to show how to teach focus and then moved into exercises showing how to teach relevant/irrelevant using the target cup stick and when cueing other behaviors.

So how do teach our dogs what is relevant and what is irrelevant?

This is really about teaching the skill to the dog: Do I pay attention or not? This is something you can set up in your training sessions by introducing new stimuli in a thoughtful manner and allowing the dog to learn to make the decision relevant/irrelevant?

She had a nice example from Simon Gadbois. One of his projects is training dogs to identify ribbon snakes by their scent instead of their appearance. Ribbon snakes look very similar to garter snakes, but they smell different, so teaching the dogs to recognize and discriminate between the scent of ribbon snakes and other snakes is a more effective way to teach the dogs to find them.

He starts by putting out two samples that have snake scent on them. Everything is identical, but the sample with the garter snake scent has been frozen, which makes it less strong. He teaches the dog to choose the ribbon snake scent, which is easier to find because it is stronger. But, over the course of the training, the sample with the garter snake scent gets stronger (or “fades in”) as the sample defrosts. After 8 hours, the scents are of equivalent strength, but because the dog initially learned that the smell of the garter snake was irrelevant, it has no confusion when the smells are of equal strength. The dogs continue to find and be reinforced for choosing the ribbon snake smell, even as the garter snake smell gets stronger.

This should give you some ideas for how we can teach dogs what is relevant and irrelevant. You want to start with something very small or at a distance, and teach the dog that it is irrelevant. Then, as the stimulus becomes stronger, the dog should assess but continue to regard the stimuli as irrelevant. Think about how you could do this with a person with a toy. Start at a distance and let the dog decide if it is relevant or irrelevant.

Exercises:  Target stick with Cup work:

The cup work contains a nice example of teaching relevant vs. irrelevant. Is left arm movement relevant?

In the first part of the session, she carefully put food in the cup while the dog was eating. But after a few clicks and treats, she started to move her left arm around a bit at random moments during the training. Once he was ignoring it, she could then start to use her left arm to put food in the cup and the dog was not distracted by it. It was a good example of first teaching him that her left hand movement did not require his attention, and then being able to use it without distracting him. This gave her a little more flexibility in the timing of when she reloaded the cup.

Is a person walking relevant?

Once the dog was following the cup, she had a person start walking slowly around the outside of his working space. Once he was ignoring the person on the outside, she had the person change her pattern so that she was sometimes walking through his workspace and the dog continued to ignore her. If the dog was distracted by the person, she had the person stop and wait until the dog was done assessing and then walk off when he focused on the cup again.

Exercises with “trained” behaviors: what is relevant?

Most of the rest of the dogs did a different exercise. She had the owner choose a behavior that the dog knew well and she structured the exercise so that there was an opportunity for the owner to change something about their behavior between food delivery and the next cue. She suggested the owner start with a minor change that was a familiar action. Pick simple things like adjusting your glasses, touching your face, fixing your shirt, etc… Then she or he could do more obvious things like picking up one foot, waving an arm, and larger movements like that. The exercise looked like this:

  • Trainer: cues “sit” or “down”
  • Dog: sits or lies down
  • Trainer: clicks and tosses food away toward the edge of the training area
  • Dog: goes to get food
  • Trainer: adds in “something new” so she is doing it as the dog is returning.
  • Dog: if dog returns as it normally would (ignores irrelevant detail), the owner cues again and repeats, doing something else the next time. If the dog returns and seems unsure, the owner stops the extra movement, waits for the dog to orient to her and cues the behavior again. Kay did mention it was better to cue the behavior than to click the dog for re-orienting to you after the hesitation.

She pointed out that the trainer’s goal is to TEACH the dog, not to TEST the dog. You really want to set it up so the dog is successful and you choose small changes so that the dog can quickly go “oh, irrelevant, what are we doing next?”

She worked with several dog/trainer pairs on this exercise and it was interesting to see how some types of movements were more distracting than others and also how some dogs would offer default behaviors if they got confused. This is not a bad thing, but indicates that there needs to be more work done on teaching the dogs about relevant and irrelevant details.

Here are a few other notes from the session:

• At the beginning of one session, she had the trainer put her treats away on a side table. This changed the dog’s focus from the trainer to the treat location. It actually allowed the trainer to be able to tell when the dog was ready to focus on her and not on the food.

• Remember you are teaching the dog what is relevant/irrelevant under these conditions. Some of this will transfer to other conditions, but you can’t count on it.

• Some dogs were confused when the trainer added hand motions that were similar to those used in food delivery. Kay said that hand cues need to be very distinct from food delivery. If you’re going to use a hand signal, it has to start from the shoulder.

• There was one little dog who was being cued to down and he was clearly using the edge of the marked area as a mat. He would down and then scoot into location. Rather than continue to cue “down,” she had the trainer use her “mat” cue since that was what he was actually doing.

• You can practice this on your own without the dog before you start your training. Think of what you want to teach the dog and plan the order (start with small easy irrelevant details) and how you will do them. This can be harder than it seems and is yet another example of the gap that often exists between understanding and doing.

• Can something that has been determined to be irrelevant now become relevant? That is one question Kay has been asking herself. If the answer is “no,” then it explains why it is better to add a new cue if the dog has learned to ignore an existing cue because it was not added at the right point or in the correct way to be meaningful.

A few reflections:

I first encountered the idea of irrelevant details (by another name) when I was learning about cues. It was not uncommon to find out that the cue I had been using was not exactly the same as the cue the horse was using. Often there were environmental components that I had not considered and when I changed them, my cue no longer “worked.”  It was quite an effort for me to learn to train so that I varied those details that I didn’t want to have attached to the cue.  But it has made a difference in my horse’s understanding of what each cue means and what information they need to consider before responding.

One aspect of relevant/irrelevant details that I think is worth consideration, especially for horse people, is how you handle novel objects in the environment. When my horses were younger, I encouraged them to stop, look and approach novel objects and see them as potential sources of reinforcement. This worked great but I found that sometimes I did want the horse to ignore the object (accept it as irrelevant) and that by encouraging interaction, I was making any new object relevant.

So over time my thinking has shifted a bit and there are times when I do allow interaction, but there are also times when I set up my training to make it clear that I do not consider the novel object relevant and neither of us need to pay attention to it.   I still allow assessment, but I try to set up my training so that I can communicate to the horse that we are going to continue what we were doing by providing clear information about what I want her to do.

Categories: Uncategorized

Tags: , , ,

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s