These are my notes from the 6th Art and Science of Animal Training Conference, held in March 2014 at the University of North Texas. Visit this page for more conference notes.
Kay Laurence’s lecture was titled “Examining 1% of the learning cycle that can add 100% success to the result.” It was all about the little details during training that trainers often do not spend a whole lot of time thinking about. Little details, however, that can sometimes make a big difference during training!
In her talk, Kay Laurence spoke about the learning cycle, meaning the whole interaction between animal and trainer. This includes the antecedents and cues (what happens before the behavior), the animal’s behavior, and the consequences (what happens after the behavior). Kay refers to this as a cycle because it repeats. After the animal consumes the reward, the trainer gives the next cue and the process begins again.
It pays off to take the time to analyze and assess what you are doing during a training session. For instance, if we consider a dog sitting, there are many different ways that the dog can get into the sit position, quite a few different ways to actually “sit,” many different places to sit, and many different things for the rest of the dog’s body to be doing. A trainer can get into trouble by not being aware of all of these different variations and not noticing what exactly the animal is actually doing.
Kay also reminded us that it is so important to look at the bits between each of these three main pieces (cue, behavior, consequence). What happens in the interval after the cue is given and before the animal does the behavior? What about after the behavior, but before the animal consumes the reward? And how about what happens after the reward is delivered, but before the next cue is given?
Kay’s talk largely focused on these in between bits. Trainers usually focus on the cues, behaviors, and consequences. But, focusing on the pieces in between can really help you understand why certain things are happening during a training session and what you can do to improve your training.
How do you give a dog a treat?
A large portion of Kay’s talk focused on the moments during training that happen after the animal does the behavior, but before the animal consumes the reward. That is, the reward delivery phase.
I think most trainers spend time thinking about the animal’s behavior and the cues the trainer will use during training. However, I think trainers often spend very little time considering how they will deliver the reward and whether this is the best method for the behavior that is being trained. Different reward delivery strategies can make it much easier (or much harder) to train certain behaviors.
Kay Laurence discussed several different patterns of reward delivery and how she uses them in her training. While discussing each, she touched on the mechanics of delivering the reinforcer, suitable reinforcers for this delivery pattern, what this delivery pattern was best for teaching, how the click functioned during the pattern, and how the reinforcer delivery might interrupt or affect the behavior. Here are some details about each of the patterns Kay discussed.
In position. During this reward delivery strategy, the trainer clicks and the dog remains in position while the treat is delivered. For Kay Laurence’s dogs, she clicks and then leans slightly toward the dog to communicate that the dog should remain in position while the reward is delivered. This can be used with treats or physical/social contact as the reward and is a great method when working on passive behaviors, opening behaviors, waiting or stillness, or husbandry behaviors. Since the dog stays in position, the delivery of the reward does not interrupt the behavior.
Go and get it. During this delivery strategy, the trainer flicks or tosses a treat or points to a station where the dog can go to collect a treat. Since the dog must go find and collect the reward, this works best with visible treats or with toys. Don’t use a treat the same color as the flooring for this! This strategy does interrupt the behavior. However, this can be useful for building up energy and for setting up the dog so he can do the behavior again. Kay commented that this is also a good strategy to measure whether the dog wants to keep coming back to participate in training.
Deliberate placement. In this strategy, the treat is used to move the dog or reset the dog into a certain position. That is, after the click the treat is used to maneuver the dog into a different position. When the dog reaches that position, the hand opens up, inviting the dog to take the treat. This can be used with both treats and social contact. This method is great when shaping micromovements or when resetting the dog to a certain position to prepare for the next behavior.
Catch. During this reward delivery strategy, the treat is tossed to the dog, so that the dog can catch the treat in his mouth. This is best done with large, easy to see treats and can be used when working on maintaining behaviors when at a distance from the dog. Kay did comment that some dogs never seem to quite master the skill of catching treats.
Kitchen! During this delivery strategy, the dog and trainer go together to a reward station, often with a massive change in the environment (such as going across the room or to a different room). This has several very different applications. For example, Kay talked about teaching a “Kitchen” cue, which would mean to the dog “Let’s go to the kitchen and get a treat.” This can work well if you run out of treats while training or if the dog does something wonderful when you do not have any treats with you. This could also be used when outdoors to move the dog away from an unsafe location. Finally, reward stations are great for active behaviors when the dog needs to maintain travel in a certain direction (such as getting the dog to keep going in a straight line after completing a jump).
Before you start
Kay Laurence cautioned that before starting any training session, it is very important to consider the reward delivery strategy that will be most suitable for the behavior you are trying to teach. Don’t pick a delivery strategy just because “it is the one you always use.”
As Kay explained in her talk, how we deliver the reward will set up the animal for the start of the next behavior. The delivery of the reward should help position the animal so that the animal is in the optimal position to do the next behavior. The delivery of the reinforcer can also affect the energy of the next behavior. This is important if you are trying to add more (or less) energy to a behavior or to just maintain the current level of energy.
Kay Laurence closed her talk by stating that a small one percent change to your training procedure can change the whole experience for the animal. I’m really interested to spend some more time thinking about Kay’s talk and in particular different reward delivery strategies and how using different strategies in certain places could improve my training.
For example, when training my rats, I almost always use delivery stations, where the rat comes to a certain place to collect the treat. I think I could probably use some of the other strategies to improve my training (in particular, deliberate placement). Now, I wonder if I could teach a rat to catch a tossed treat? Probably not, as rats have quite poor eye sight, but it might be fun to try!
What methods do you use to deliver rewards during training? Do you have any methods you use that are not listed here? How do you decide which method to use? Which of these might you want to explore further in your training?