The Cognitive Dog:Savant or Slacker • Class 14: Observational Learning
Agenda • Final paper • Next week • Carolyn on learning theory • Bruce on observational learning theory
The brain trust • Lupe & Willow
The brain trust • Maggie & Cooper
The brain trust • Rosito & Pedro
The Cognitive DogClass 13 • ABC’S OF LEARNING THEORY • How Dogs Learn – The Basics
How Dogs Learn • CLASSICAL CONDITIONING • Also known as Pavlovian, Associative and Respondent • OPERANT CONDITIONING • What we know as “training”
How Dogs Learn • CLASSICAL CONDITIONING • Pavlov’s Dogs • Relates to reflexes not to voluntary behaviors • Happening all the time • Clicker and food is a classical association • Teaching a cue is a classical association
How Dogs Learn • OPERANT CONDITIONING • Definition (from How Dog’s Learn Burch & Bailey, 1999): • “The part of science of behavior • that explains the functional relationship • between environmental events and • behavior. It is a key component in • explaining how all organisms • (including dogs) learn.” In simple • terms what happens in the environment • will affect what a dog will do in the future. • If a dog receives a cookie for sitting he will • be more likely to sit again. If a bee stings a • dog when he sits he is less likely to sit.
How Dogs Learn • OPERANT CONDITIONING • REINFORCEMENT: • Strengthens behaviors • . • PUNISHMENT: • Weakens behaviors
How Dogs Learn • OPERANT CONDITIONING • REINFORCEMENT: • Strengthens behaviors • Is anything that will increase the likelihood of a behavior to happen in the future • To be effective it should closely follow the behavior so an association is made
How Dogs Learn • OPERANT CONDITIONING • PUNISHMENT: • Weakens behaviors • Punishment decreases the likelihood of a behavior to be repeated in the future • To be effective it should closely follow the behavior • Has potential fallout; fear, anxiety
How Dogs Learn • REINFORCEMENT in dog trainingStrengthens behaviors • Positive Reinforcement • Add good • As perceived by the dog • Negative Reinforcement • Remove bad • As perceived by the dog
How Dogs Learn • REINFORCEMENT in dog training • Positive Reinforcement • Add good - As perceived • by the dog • Examples: Food, toys, exercise, • play & petting (not all dogs)
How Dogs Learn • REINFORCEMENT in dog training • Negative Reinforcement • Remove bad, As perceived • by the dog • Examples: Stop pulling on a leash, stop electronic collar; spray or shock, stop ear pinch
How Dogs Learn • REINFORCEMENT in dog training • Positive Reinforcement • Primary Reinforcer • Secondary Reinforcer • Negative Reinforcement • Primary Reinforcer • Secondary Reinforcer
How Dogs Learn • REINFORCEMENT in dog training • Positive & Negative Reinforcement • Primary • An Unconditioned Reinforcer. A primary reinforcer will increase the likelihood that a behavior will be repeated in the future when presented immediately following a behavior. Related to basic needs, food, drink, some touch. • Secondary • A Conditioned Reinforcer. It is something that by itself does not mean anything to the dog. Butwhen paired with a primary reinforcer the secondary will then strengthen behaviors. For dogs these can include, praise, clicker, petting, verbal reprimand or sound. This pairing is classical conditioning. • REINFORCEMENT INCREASES BEHAVIOR
How Dogs Learn • Punishment in dog training: • Punishment decreases behavior • Positive Punishment • add something bad • Negative Punishment • take away something good • The good or bad must be the dogs perception
How Dogs Learn • Punishment in dog training: • Positive Punishment • Add something thing bad • Examples: electric shock, leash correction, spray bottle or collar, throwing objects, hitting
How Dogs Learn • Punishment in dog training: • Negative Punishment • Remove something good • Examples: Time out in crate = removes you or the fun, Take dog away from fun event such as agility run for broken stay
How Dogs Learn • Punishment in dog training: • Positive Punishment • Primary Punisher • Secondary Punisher • Negative Punishment • Primary Punisher • Secondary Punisher
How Dogs Learn • Punishment in dog training: • Positive & Negative Punishment • Primary • An Unconditioned Punisher. The dog does not need prior experience for it to be perceived as a punisher. Extreme heat or cold, shock, pinch, citronella spray, hitting are all examples of primary • Secondary • A Conditioned Punisher. It is something that by itself does not mean anything to the dog. Butwhen paired with a primary punisher the secondary will then weaken behaviors. Verbal or sound. This pairing is classical conditioning. • PUNISHMENT DECREASES BEHAVIOR • But there is possible increased anxiety and fallout
How Dogs Learn • OPERANT CONDITIONING • ReinforcementPunishment • PositivePrimary: Food Shock Secondary: Click Warning Sound • Negative Primary: Remove Shock Remove Food • Secondary: Remove Verbal Remove Verbal
How Dogs Learn • finally…The ABC's of learning • A= Antecedent or A Signal (cue) • B= Behavior (action) • C= Consequence (payoff) • Good Consequence = Reinforcement • Bad Consequence = Punishment
How Dogs Learn • The ABC's of learning • A= Antecedent or A Signal • Cues • Verbal: words, sounds • Visual: hand, body, motion • Environmental: agility obstacles • Think about the dog youhave when making choices for cues; visual or sound stimulation stronger?
How Dogs Learn • The ABC's of learning • B= Behavior • Action or duration of non action • Voluntary action: sit, down, lift paw • Non action: stay in position • Behaviors that are offered by the animal appear to be learned more quickly vs. guiding or continuous luring
How Dogs Learn • The ABC's of learning • C= Consequence • Payoff • Good Consequence = Reinforcement, R+ • Increases likelihood of behavior • Bad Consequence = Punishment • Decreases likelihood of behavior
How Dogs Learn • Once learned, Good Consequences = • Behaviors Happen: • Human Example: • A = Antecedent (cue): Alarm goes off, ugh… • B = Behavior: Get up and go to work • C = Consequence: Paycheck, yahoo! • Canine Example: A = Antecedent (cue): Verbal “Sit” • B = Behavior: Dog puts butt on the floor • C = Consequence: Dog gets to go out to play!
How Dogs Learn • “Natural” Progression of A-B-C • Begging: • A - Owner eating at table • B - Dog Begs • C - Owner gives a piece of food • Think the begging will be repeated?!
How Dogs Learn • Training Steps: A new alphabet • B-C-A • Maximizes learning & the quality of the final behavior performance • B = Behavior: Get it • C = Consequence: Give it • A = Antecedent or A Signal:Add after behavior is strong, insert before behavior
How Dogs Learn • Training Steps • B = Behavior: • Get the dog to perform the desired behavior • Capture: R+ dog doing behavior • Shape: R+ parts of the behavior working closer to the end behavior • Lure: Use of food or toy as a guide to where you want the dog then R+
How Dogs Learn • Training Steps • C = Consequence: • R+ • When the dog does the behavior- Capturing • When the dog does part of the behavior- Shaping • As the dog follows the food into position or partial position - Luring
How Dogs Learn • Training Steps • A = Antecedent or Cue: • Add after the dog becomes good at the behavior, cue will pair with the end behavior performance (usually better than beginning) • Cue is presented just before the dog does the behavior = Classical Conditioning • R+ as usual
How Dogs Learn • EXAMPLE OF • CAPTURING – SHAPING – LURING • Down – • Capture: Sit in a small quiet room R+ as dog lies down on his own • Shape: R+ as dog lowers head, then bends elbows, then lowers back end etc. • Lure: Guide dog into position with food or toy desired by the dog, R+ in position
How Dogs Learn • BUT KEEP IN MIND…
How Dogs Learn • Emotions & Motivation impact Learning • Take all we have studied so far into consideration… • Emotions • Hardwiring - Motivation • Past Experiences • Lack of Experiences • Keep in mind all of this and the immediate environment have a big influence on a learners ability to learn
How Dogs Learn • THE LEARNERS …that have taught us….
Things to note... • The other animal is just doing their thing, their role as “demonstrator/teacher” is purely a side-effect. • Think of wolf pups following mom & dad • Their presence/interaction/response provides, in effect, a focus of attention that simplifies the learning problem for the observer. • Associative learning is probably sufficient given this scaffolding
Local enhancement: “I think I’ll hang out with Harry. Hmmm, what’s that” Stimulus enhancement: “Hey what’s Harry fooling with. That looks tasty” Observational Conditioning: “Yikes, what is Harry reacting to, I guess I should be scared too” Three simple types of social learning • Just because its simple, doesn’t make it any less useful
Uncommon forms of observational learning Goal Emulation Imitation Emery, N. and N. Clayton (2005). Animal Cognition. The Behavior of Animals: mechanisms, function and evolution. J. Bolhuis and L. Giraldeau. London, Blackwell Publishing.
Goal Emulation... b) Achieve same goal but use a different action a) Observe another animal achieve a goal by performing a specific action The difference in this case is that the person and the monkey orient the rake differently Emery, N. and N. Clayton (2005). Animal Cognition. The Behavior of Animals: mechanisms, function and evolution. J. Bolhuis and L. Giraldeau. London, Blackwell Publishing.
Imitation... b) Achieve same goal by performing the same specific & novel sequence of actions a) Observe another animal achieve a goal by performing a specific & novel sequence of actions This type of observational learning seems to occur rarely in other species than humans Emery, N. and N. Clayton (2005). Animal Cognition. The Behavior of Animals: mechanisms, function and evolution. J. Bolhuis and L. Giraldeau. London, Blackwell Publishing.
Why? • Why are goal emulation and imitation apparently so rare in species other than humans and maybe other apes... • May require a type of cognitive machinery (little or big) not found in other species... • The cost/benefit may low (i.e., the common forms work well enough given built-in behavioral structure & context) • Just haven’t looked hard enough...
What do our slacker friends do? • Common forms of observational learning... • Local enhancement • Stimulus enhancement • Observational conditioning • Less common forms • Goal emulation (???)
Big idea... • Test for evidence of observational learning in Miniature Dachshund pups of varying ages... • Had some pups watch another pup retrieve ‘inaccessible’ food by pulling on a string that was attached to the food • Compared time it took naive pups to retrieve food by pulling on string with the time it took pups who had experience watching other pups do it.
More details... • 4 litters of doxie pups, 1 from each litter chosen as demonstrator for rest of littermates • 3 days of acclimatization (???) • 5 trials per day in which demonstrator learned how to retrieve food, and observers watched • After 15 trials, observers were given chance to perform task
Results from Adler... • The observers generally retrieved the food substantially faster on the first three trials than their demonstrator • Particularly strong effect on 60 and 38 day-old litters • Evidence that they benefited from watching demonstrator