So Many Choices: for trainers & for dogs

So Many Choices:

for trainers & for dogs

Kathy Sdao, MA, ACAAB

www.kathysdao.com

Choices

Trainer’s choices include: What will you teach exactly ? How will you “get behavior”? Will you use R+, R-, P+, P-, Classical Conditioning?

Combo? What reinforcers? Kept where? Delivered how? How often? What are criteria for success? How will you change

criteria? How will you set up the training environment? How will you structure a training session? When will you add a cue? What cue? How will you judge if the behavior is fluent?

Trainer’s influences

Habit & practice your own reinforcement history

Role models in person, on TV, etc.

Education conferences, seminars, books, DVDs, etc.

Personal values*

You aren’t the only one making choices.

Dogs face hundreds of “choice points” daily.

Reinforce as many correct choices as possible.

FrequentlyPreciselyGenerouslyReally

Be a “choice architect.”

Manipulate environment so when dog is at choice point, she’ll choose correctly 80-90% of the time.

Jungle-path analogy (Jean Donaldson, The Culture Clash, p 51)

▲

Behavior humans prefer: Make easy

Behavior dog prefers: Make hard

SMR

Reinforce dogs’ choice points.

Set up the situation cleverly, then observe.

Pay attention! Notice correct choices.

Spend more energy watching the dog than luring, prompting or physically manipulating the dog.

This allows the dog to make free choices, without becoming dependent on your help.

Explain training to a novice in 30 seconds

Ann & Boomer

What’s

essential? What’s important? What’s irrelevant?

Get SMART

See

Mark

And

Reward/Reinforce

Training

Back-chaining a behavior chain (e.g., soda fetch)

First, train each link to fluency & give each a cue.1 2 3 4

Cue last behavior several times; reinforce well.“4” R+

Cue preceding behavior, then final behavior; reinforce. Repeat until fluent.

“3” – “4” R+

Next add the preceding behavior, etc…“2” – “3” – “4” R+

Back-chaining core skills for the trainer

(Dog moves) Trainer Reinforces

(Dog moves) Trainer Marks Reinforces

(Dog moves) Trainer Sees Marks Reinforces

R = Reinforcement

Strive to create dozens of reinforcers for each animal you train.

“Our job is to maximize the efficiency of positive reinforcement.” BF Skinner

To be a reinforcer, a consequence must…

1. follow an action = be contingent on a response

Time

2. cause the action it follows to be repeated or occur more often

Function

“Should Kids be Bribed to do Well in School?”Time magazine, April 18, 2010, by Amanda Ripley

The more reinforcers, the better

Pavlovian conditioning: creates new conditioned reinforcers

CS + US B (reflex)

The Premack Principle:dog’s distractions = potent reinforcers

“reinforcement is the opportunity to exchange

a less probable activity for a more probable one”

Another source of R+

Cue must be familiar; behavior must be fluent Cue must have been trained with R+

Cues = conditioned reinforcers

Tertiary PrimarySecondary

“The Precious Cue”

Good news!Opportunity for dozens of additional

reinforcers

Essential for training links in behavior chains

Bad news! Don’t give cues simultaneous with bad behavior

Worst for dogs with few other ways to earn R+

Reinforcement maintains behavior.

“You have to floss only the teeth you want to keep.”

Your choice:1) positive reinforcement

= the addition of treats; satisfactionor

2) negative reinforcement

= the removal of aversives; relief

The opposite of reinforcement is…?

… no reinforcement. (Not punishment.)

Differential reinforcement requires you to avoid reinforcing errors. “refusal” to respond to a cue incorrect response to a cue imprecise movement (e.g., weave-pole entry)

“bad” behavior

M = Mark

Why bother?Because for animals, figuring out what behaviors “work” is harder if the reward is both:

1) the reinforcement &

2) the info about which behavior was correct

TSA analogy

Clicker functions

Conditioned / Secondary reinforcerStrengthens behavior

Cue or Sd

The occasion when going to the feeder is reinforced

Event markerPinpoints the desired behavior

Bridge Bridges time between behavior & treat Promise = reinforcement is coming

Clicks are the fulcrum

“There are two sides to the click: what happens before, and what happens after. What happens immediately before the click is a behavior the trainer would like to strengthen. What happens immediately after is an event the animal would like strengthen, such as receiving food. The click unites these two desires.”

Alexandra Kurland

www.theclickercenter.com

What weakens your marker?

It’s redundant/non-informative

It’s poisoned

It’s infected with too many anxious situations

You require more behavior after marks

It marks non-behavior

Marking precisely…

You can’t take a click back

Timing improves if you can predict behavior

This is a function of seeing well

Mean simple reaction times for college students

= ~0.19 sec for visual stimuli;

= ~0.16 sec for sound stimuli

www.humanbenchmark.com/tests/reactiontime/index.php

S = See

Seeing is influenced by…

Preconceptions & labels“Who would you be without your story?” Byron

Katie

Judgments & analysis“Don’t think, just look.” Ludwig Wittgenstein

Talking & prompting

The audience effect

Seeing practice

“To see what is in front of one's nose needs a constant struggle.” George Orwell, 1968

1) “Surprising Studies of Visual Awareness” (DVD)

Simons, Daniel (2008)

www.viscog.com www.theinvisiblegorilla.com

2) Watch videos of animals moving, in slow motion

What precedes the “S”?

Your decision: Search pattern for a specific clickable behavior

Your attention: Human brain processes one-half of one millionth of sensory data it receives.

Consciously choose to filter in correct behaviors.

Your set-up: Strongly influences the dog’s behavior

S = Set-up: the other “S”

Control the environment not the dog

Limit dog’s activities & access

Much good training is undone by allowing dog to rehearse bad behavior outside training sessions.

dog barks out front window at passersby dog lunges at another dog on hurried walk in busy

park

Reactive dogs often need to be confined, especially early in retraining program.

Careful set-up pays off.

Ideally, distractions becomes cues for correct behavior

Morton’s recall video Rusty & Charlie responding to a “mock knock” In my car, Effie looking at me when she sees a

dog

Best if distractions are added late in training sequence, as “new cues”

Thank you!

Getting behavior

Trainers want to change behaviors.

But we can’t manipulate behavior directly.

We can manipulate events immediately before a behavior: antecedents

Or we can manipulate events immediately after a behavior: consequences

A B C

Ten ways to “get behavior”:

1) Physical pressure (“molding”)2) Prompting3) Luring 4) Targeting5) Capturing6) Shaping7) Classical conditioning (e.g., music trick)

8) Removal of inhibitors9) Modeling/mimicry (for primates especially)

10) Verbal instructions (for humans who share language)

More seeing practice

“Spin”

Nick:

Effie:

Garth: