Notes from ClickerExpo (Friday)
For those who want a short and clear summary of what Clicker Training is, here’s one good article.
Sarah our trainer was there and I met some really nice people. To my Doggie Drawings clients and Boogie’s blog readers, you were probably there too and if I had known what you look like, I would’ve said hello!
It was so great to be at an event with dogs present. I didn’t bring Boogie because, as you know… he’s a reactive fella, and he would’ve been insanely bored during the seminars :)
Many of the seminars that I was interested in were scheduled at the same times so it was hard to decide which one to go to. If time and money were no object (and if I didn’t miss Boogie too much) I would have registered for the full three days and spent the weekend at Newport Beach.
It was such a blast to see and hear Karen Pryor in person. I loved her book “Reaching The Animal Mind” which is what inspired me to learn about clicker training in the first place.
Karen showed a video of a clicker-trained rhino doing a SIT, DOWN and ROLL OVER (!)- which was a great example of training without the use of physical contact or force …not something that you can do anyway with animals way larger and deadlier than the average reactive dog. The thing that fascinates me so much about Clicker Training is the claim that it works for ALL animals… wild ones, tame ones, zoo animals, farm animals, and even your regular household fish.
This kit was sold at the ClickerExpo store.
Seminar 1: Cecelie Koste – CLICKERTRAINING 101
As much as I wanted to sit in on Ken Ramirez’s talk on “Aggression” which everyone was raving about afterwards, I am really glad I chose Cecelie Koste’s Foundation Level seminar because…wow, I finally understand the language that dog trainers use!!! Cecelie was very clear in explaining all the concepts of Operant Conditioning, and how these apply to the Clicker Training methodology. I loved the video examples and I took a ton of notes, which I will eventually re-share on this blog as drawings.
First, Cecelie explained the difference between Classical Conditioning (developed by Pavlov, for changing emotions) and Operant Conditioning (developed by Skinner for changing behaviors). The latter is what Clicker Training is based on.
There are some Operant Conditioning concepts that I have always struggled with, and I know that I am not alone. Whenever trainers mention the terms Positive Reinforcement, Positive Punishment, Negative Reinforcement, and Negative Punishment as the Four Quadrants of Operant Conditioning, it is SO HARD not to think of “Positive” as meaning “Good” and “Negative” as meaning “Bad”, when these are really abstract terms that refer to the act of “Adding something”(+) or “Subtracting something”(-) in the training process. Similarly, the term “Punishment” means “decreasing a behavior” and not the act of doing something nasty.
A lightbulb went off in my head yesterday morning. The way I understand it now is that these consequences WORK IN TANDEM, AS PAIRS.
The way Cecelie explained it, Positive Reinforcement works in tandem with Negative Punishment, and Positive Punishment works in tandem with Negative Reinforcement. Personally, I don’t get why they still have to be thought of as “quadrants”… It’s easier to remember them as twosomes.
Positive (+) Reinforcement = Adding GOOD STUFF that your dog will work for. Treats, praise, petting, toys, sniffing, etc. to increase a behavior.
Negative (-) Punishment = Removing or Delaying the GOOD STUFF (treats, praise, petting etc) to decrease a behavior. Like a “Time Out”
Positive (+) Punishment = Adding BAD STUFF like shocks, pinches, leash pops, scolding etc to decrease a behavior.
Negative (-) Reinforcement: = Delaying or Escaping the BAD STUFF (shocks, leash pops, scolding etc) to increase a behavior. Like a “Warning”
Furthermore, here’s what I picked up — there are always TWO BEHAVIORS going on at the same time if one is being reinforced and one is being punished.
[click on picture to see it larger]
Even though this all seems like common sense from my own varied training experiences with Boogie, it is interesting to get a scientific breakdown.
With the +R/-P training - the dog walks on loose leash because he gets to move forward and get paid with food. With the +P/-R training – the dog walks on loose leash because he is avoiding having his neck choked. The first pair of consequences +R/-P is what Clicker Training is about and has been proven to be much more effective than +P/-R because the dog remains happy, un-stressed, and eager to learn.
Cecelie talked about the concepts of Extinction and Variable Schedule of Reinforcement and how it contributes to Shaping new behaviors. Other concepts covered: PRIMARY & SECONDARY REINFORCERS, HULL’s DRIVE THEORY, THE PREMACK PRINCIPLE, CUES, why BACKCHAINING is more effective than FORWARD CHAINING (Boogie and I have done some backchaining work with Sarah – scroll down). Like I said, more food for thought –> more ideas for illustrations! :)
Seminar 2: Kathy Sdao – On SHAPING
Kathy Sdao is such an amazing, enthusiastic, passionate and entertaining speaker! Never a dull moment. I have watched some of her DVD seminars but this live presentation was so much better. Quotes and notes to follow.
The ABC of training: Antecendent –> Behavior –> Consequence
“We can’t manipulate behavior directly. We can manipulate events just before (Antecedents) and just after (Consequences) a behavior.”
“The most fundamental law of behavior is that Consequences drive Behavior”. Which I guess is like saying that we work in order to get paid, not because we were asked to. Or that Boogie barks at the mailman not because of anything the mailman did, but because the barking *causes* the mailman to go away. I can see that this applies to the BAT work that Boogie and I have been doing, where the “consequences” = “functional rewards”.
99% of “Stubbornness” is trained. Clicker training gives the animal a sense of AGENCY. “I made her click!”. It teaches animals that they have control, they can change their world.
“Shaping is the heart and soul of clicker training”.
Prerequisites for Shaping behavior:
1. Clear goal behavior, visually and verbally.
2. What is the 1st criterion that is clickable? Choose something that he already does.
3. Ability to SEE.
4. Click-Treat mechanical skills… click precisely, and click often.
5. Click MOVEMENT. “The Dead Dog Rule”: don’t click non-behavior (eg, silence or stillness)
6. A hungry and confident animal
7. Very small delicious treats
Kathy demonstrated a Shaping/targeting exercise with a dog in the audience (she placed a bottle down on the floor in front of the dog) and it was kinda funny to see this little pup Olivia do some of the same stuff that Boogie does. Several times, Olivia sat down and waited for the treat instead of moving forward to offer new behaviors. (Yay! My dog is not a freak!) I loved that Kathy explained what her own mistakes were in this exercise, which shows that even a veteran master animal trainer isn’t perfect. :)
Seminar 3: Kay Laurence – On MICROSHAPING
I had registered myself for a Learning Labs Workshop with Kathy Sdao to see real live dogs being shaped but at the last minute I changed my mind and attended Kay Laurence’s seminar instead… mainly because I was curious to get another perspective on the same subject.
Kay was very impressive and funny (very dry British humor) and had very strong opinions on the Ethics of dog training.
“Training should be Learner-Centric, not Behavior-Centric.” ie, just because you can train it doesn’t mean you should. Kay Laurence is not a fan of YouTube videos of dogs trained to walk on their two front legs. “Shaping is working with a natural movement, not teaching an unnatural one”. (I wonder if she is referring to Neo and Tuxedo…) Kay believes that a lot of training is done for the trainer’s ego or with no respect for the animal. One example being female animals who are trained to stand perfectly still for breeding purposes… :(
She is also not a fan of the “101 things with a box” game or “freeshaping” because it lessens the effectiveness of teaching precise behaviors and causes confusion for the dogs.
A lot of the stuff she talked about was perhaps too advanced for me… Most of the technical details went way over my head and perhaps are more relevant to dog owners who do Obedience competitions and have to get very precise “beautiful” behaviors from their dogs.
“Shaping is 90% planning and 10% training”. *Yep, the planning part is the hardest for me!
It is better to do 10 x 1 minute sessions, than 1 x 10 minute session. There has to be a clear agenda for each session.
You want a high 95% success rate. “Success is measured by the rate of reinforcement not the speed at which the final behavior is achieved”. The dog is reinforced frequently, for tiny increments of improvement.
From what I understand, Kay’s “Microshaping” method is not so much about shaping or clicking the outcomes. (eg, dog touches ball) It focuses on shaping the micromovements – accurate, refined muscle movements (dog lifts paw) so that the dog is very clear on what he is doing and you get high quality behaviors. She showed a few videos where a dog was clicked for repeatedly moving his paw to a specific position. And then she added the target object in the same place where his paw would go and clicked the dog for putting his paw on the object. A different protocol from the one where you first present the target object in front of the dog, then click his movements towards it.. as Kathy Sdao had done. (Sarah and I have done something similar with Boogie)
She also showed videos of a dog standing on dome-things (targets) and pointed out the different muscles moving on the dog’s body when she changed the positions of the targets. Whoa. I guess one would need a really expert understanding of a dog’s physiology and natural movement patterns to train this way.
Kay Laurence: “Some people taste a dish and know how to cook it. I can look at a behavior and know how to teach it.”
All in all, that was a fantastic learning experience, and if money were no object next year, I would do it again!