Transcript document

Learning
What counts as learning?





A child speaking a second language with more fluency
every day
A parrot repeats a new sentence
A dog who waits by the door for you every time it hears
the school bus
A computer that has been programmed with a new rule
for solving problems
You think about a cool new fact you just heard in AP
Psychology


Learning: a relatively permanent change in behavior due to
experience
We learn most things by association


Animals learn to associate one stimulus with another
(classical conditioning)


Ex: Jaws theme song signals shark attack
Ex: sea slugs can learn that water squirts signal electric shock
Complex animals associate their own behavior with results
(operant conditioning)

Ex: dogs associate begging with being fed
Ivan Pavlov



Behaviorist
Studied digestion
Taught dogs to salivate at the sound of a buzzer by
associating the buzzer with blowing meat powder into
the dog’s mouth
Classical Conditioning




Before conditioning:
Present unconditioned
stimulus (US, aka UCS)
Give the dog food
Result: dog will salivate unconditioned response
(UR, aka UCR)




During conditioning:
Pair a neutral stimulus
with the US
Give the dog food and
sound a tone
Result: dog will salivate
because of the food



After conditioning:
Present conditioned
stimulus (CS) - tone
Result: dog will salivate
because of the tone alone
(conditioned response or
CR)
Some tricks

In any situation, replace the word “conditioned” with the
word “learned”


‘Conditioned response’ becomes ‘learned response’
The UR and the CR are usually the same - what changes
is the stimulus

Dog salivates for food and tone
Adaptive value

Natural examples of conditioning prepare us for events




Lightening warns about thunder
Smell of cooking alerts us to dinner
Helps animals avoid danger, find food, and be ready to
reproduce
Conditioning is stronger when we are aware of it
Acquisition



Learning a new stimulus-response pair for the first time
Timing is important - ideally the CS is presented first with
a very slight delay or overlap of the US
CS used to predict US
Extinction


If the CS is presented repeatedly without the US, the CR
will eventually stop happening
Dogs will unlearn that a tone signals food if food never
shows up after the tone
Spontaneous Recovery

After not being exposed to the CS for a while, CR may
occur the next time CS does (even after extinction)
Generalization


Animals may respond to stimuli that are similar to the CS
(ex: tone of different pitch)
May be adaptive – ex dog abused by a man fears all men
Discrimination


Animals can also learn to distinguish between two similar
stimuli if the US is only presented after one of them
Also adaptive – small differences in stimuli can mean big
differences in safety
Cognitive processes?


Early behaviorists ignored all cognition
Modern research shows rats can learn the reliability of
the CS for predicting the US



Rats distinguish between two CS and only respond to the
more reliable one
Dogs (and people) can develop learned helplessness: if they
are repeatedly put in situations they have no control over,
they will not react when they do have control
Awareness of conditioning may diminish or enhance its
usefulness in therapy


If you know alcohol is spiked, you won’t generalize the result
to all alcohol
If you know positive associations are helpful, you may make
stronger ones
Biological effects

Conditioning is faster, stronger, and longer-lasting if it is
relevant to your survival





Taste aversions
Trauma
“a burnt child dreads the fire”
“once bitten, twice shy”
Conditioned stimuli with a natural connection to the
unconditioned stimuli they predict
Experiment to Know



John B. Watson used classical conditioning on baby “Little
Albert”
Used a frightening noise as the US and a rat as the CS.
Albert’s UR and CR was fear.
Albert generalized the CR to furry things like rabbits,
dogs, and fur coats
B. F. Skinner



Built the operant chamber (aka Skinner box)
Trained pigeons to guide missiles in WWII
Researched the principles of behavior control based on
law of effect:

Rewarded behavior is likely to happen again
Figure 6.12 A Skinner box Inside the box, the rat presses a bar for a food reward. Outside, a
measuring device (not shown here) records the animal’s accumulated responses.
© 2010 by Worth Publishers
Operant Conditioning


Behavior that operates on the environment leads to
either reinforcement or punishment
Animals learn to associate their behavior with its result
Shaping





Shaping is used to train animals to do complex behaviors
1. Start with what the animal can do
2. Reward it for doing something close to what you want
3. Require it to get closer and closer to the desired
behavior before rewarding it
Give an example of shaping.
Shaping cont.

Shaping also helps researchers discover what animals can
perceive, and even some mental processes (forming
concepts)

Ex: pigeons have been trained to distinguish between human
faces, flowers, and cars, and can sort new pictures into the right
category
Reinforcers


Anything that strengthens the behavior it follows
Positive reinforcement adds (+) a pleasant stimulus


Negative reinforcement removes (-) an unpleasant
stimulus


Ex:Doing your chores earns allowance
Ex: hitting the snooze button stops your alarm clock
NEGATIVE REINFORCEMENT IS NOT PUNISHMENT
(it actually removes punishment)
Primary and Secondary Reinforcers

Primary reinforcers are unlearned


Ex: chocolate is delicious
Secondary or conditioned reinforcers are associated
with primary reinforcers

Ex: money can be used to buy chocolate


Immediate reinforcement is necessary for rats
Humans can choose delayed reinforcement


working today to get paid tomorrow
Children who demonstrate the ability to wait for delayed
reinforcement become high-achieving adults
Reinforcement Schedules


Describe when and how often behavior is reinforced
Continuous: behavior is reinforced every time




Learning happens quickly
Extinction also happens quickly
Not common in the real world
Partial or intermittent: behavior is sometimes reinforced



Learning happens more slowly
Extinction is very difficult
Example?
Intermittent Reinforcement Schedules

Fixed-ratio schedule: behavior is reinforced after a set
number of times


Variable-ratio: behavior is reinforced after an
unpredictable number of times


Ex: slot machines
Fixed-interval: behavior is reinforced after a set amount of
time


EX: Loyal customer card - buy 10, get the 11th free
Checking the mail every day around delivery time
Variable-interval: behavior is reinforced after an
unpredictable amount of time

Ex: checking email
Figure 6.13 Intermittent reinforcement schedules Skinner’s laboratory pigeons produced these
response patterns to each of four reinforcement schedules. (Reinforcers are indicated by diagonal
marks.) For people, as for pigeons, reinforcement linked to number of responses (a ratio schedule)
produces a higher response rate than reinforcement linked to amount of time elapsed (an interval
schedule). But the predictability of the reward also matters. An unpredictable (variable) schedule
produces more consistent responding than does a predictable (fixed) schedule.
© 2010 by Worth Publishers
Punishment



Punishment decreases the frequency of a behavior
Can add an unwanted stimulus (speeding tickets) or take
away something desired (freedom while in jail)
The most effective punishment is swift and certain
Physical Punishment






Physically punished behavior is temporarily suppressed, not
forgotten
Physical punishment may teach discrimination instead of
stopping a behavior
Physical punishment can teach fear of the person and place,
not just the behavior
Physical punishment models aggression as a solution for
problems
Physical punishment works best as an occasional back-up
that is combined with reasoning
Reinforcement gets more permanent results and makes
people happier
Cognition and Operant Conditioning


Latent learning occurs in rats and people - learn by
observing or practicing, but don’t demonstrate learning
until later when it is needed
Insight is when learning seems to occur spontaneously
Intrinsic vs Extrinsic motivation



Intrinsic motivation comes from within (you do
something because you enjoy it)
Extrinsic is external (you do it because you get paid, or
because you have to)
Offering extrinsic motivation for something people enjoy
can cause them to enjoy it less
Biology and Operant Conditioning


Animals are better at learning things that are naturally
adaptive
Animals may experience instinctive drift away from the
trained behavior and toward their natural behavior
Observational Learning


AKA social learning - learning through observation and
imitation (modeling)
Works for complex creatures (most mamals and some
birds)
Mirror Neurons




Accidentaly discovered in 1991
Found in frontal lobe
Fire when acting or watching others act
May be useful for empathy, imitation, and infering the
mental state of others



Infants and children imitate the behavior of an adult
At 14 months children imitate what they see on TV
Mirror neurons may explain our reactions to movies,
books, and contagious yawns
An Experiment to Know





Albert Bandura “Bobo Doll” experiment
Experimental group of children watched an adult beat up
an inflatable doll, the control group did not
Experimental group mimicked aggressive behaviors and
invented new ones
Control exhibited little to no aggressive behavior
http://www.youtube.com/watch?v=YclZBhn40hU
We more readily imitate someone if


They are similar to us
They are rewarded for their behavior
Prosocial Effects of Modeling



Positive role models can increase prosocial behavior
Models are most effective when their words and actions
are consistent
Examples


read with children to encourage independent reading
Christians who rescued Jews in the Holocaust were more
likely to have had a parent with strong morals
Antisocial Effects of Modeling


Abusive or violent behavior can also be learned by
observation, in person or on TV
Many studies demonstrate a postive correlation between
viewing violence on TV or in video games and aggressive
behavior


From imitation (like Bobo Doll)
From desensitization (minimizes empathy for victims)