Learning
- Learning - a relatively
permanent change in knowledge, attitude, or behavior brought on by
experience. Not due to maturation.
- Three Types of Learning:
- Classical
Conditioning Learning by association.
- Operant
Conditioning Learning by consequences of behavior
- Obseravtional
Learning Learning by modeling others
Classical Conditioning – 1 stimulus comes to predict the occurrence of another stimulus
and elicits a response similar to related to response evoked by that stimulus
Classical Conditioning Terms
- Unconditioned Stimulus
(UCS or US)
- Produces/elicits a
response w/o previous or prior learning.
- Unconditioned Response
(UCR or UR)
- Unlearned response
(reflex) invariably elicited by the US
- Conditioned Stimulus
(CS)
- Previously neutral
stimulus that elicits a response after being paired with or associated
with the US.
- Conditioned Response
(CR)
- Learned response to CS
after pairing of US-CS. Usually similar to the UR.
Other Important Terms
- Acquisition
- Extinction
- CS not followed by US.
CR eventually ceases.
- Spontaneous Recovery
- After first
extinction, CS presented at a later time (rest period). CR again occurs
for a limited period of time in a weaker form.
- Stimulus Generalization
- Responds to stimuli
similar to CS.
- Stimulus Discrimination
- Can distinguish
between a new stimulus and the already learned stimulus. The new stimulus
is similar to the old CS.
- Higher Order Conditioning
- Neutral stimulus
paired with CS becomes associated with it, gains the power to elicit the
same CR (Not the same as stimulus generalization)
Temporal Relationships between CS and UCS
- Delayed - CS
precedes the US
and still there when US is presented (tone is presented and remains on
while food is presented). Best for conditioning
- Trace - CS
presented first but ends before the US
occurs (taste aversion)
- Simultaneous -
CS and US presented at the same time
- Backward - CS
follows the US
(tone follows the food). Worst for conditioning
Application of Classical Conditioning:
Fears - The case of Little Albert Watson & Raynor
(1920)
- Can fears be acquired
through classical conditioning.?
- When Little Albert
touched a white rat, Watson made a loud noise behind Albert’s
head.
- Albert began to cry
- Soon, Albert showed fear
of the rat.
- His fear generalized
to other, similar. objects: dog, rabbit, fur
coat, Watson in beard.
Biological Constraints on Classical Conditioning
- Can any
association be learned? Humans more easily conditioned to fear stimuli
harmful to well-being or related to survival through evolution (snakes,
heights)
- Conditioned taste
aversion
- CS
- Sweet water, flash of
light, clicking sound.
- US
- For first
group, electric shock which produces pain.
- RESULTS: Feared
light & click, but not sweet water.
- For second
group, x-rays which cause nausea.
- RESULTS: Feared sweet
water, but not light & click. CTA
Other
factors influencing classical conditioning
- How reliably the CS
predicts the US
- Number of pairings
- Intensity of the US
Operant Conditioning
- Learning that occurs when an
individual produces an active behavioral response which is followed by an
environmental stimulus, which influences the likelihood that the response
will be repeated again in the future. Consequences are manipulated to
increase or decrease behavior in the future.
Types of Reinforcers
- Primary reinforcers
- Stimuli that have
inherent value.
- Food, water,
warmth, sleep, sex
- An unlearned
reinforcement
- Secondary reinforcer
- Stimuli. that acquire
reinforcement properties by being associated with primary reinforcer
- Money, praise
- A learned reinforcer
Reinforcement
- Positive - When
the behavior occurs, a pleasant event follows.
- Get a cookie for
being good, Get paid for going to work.
- Reinforcers
must increase behavior..
- What is
reinforcing for one person may not be reinforcing for another.
- Negative - When
the behavior occurs, an unpleasant event ends.
- Not the
same as punishment!
- Negative
reinforcement increases likelihood that a behavior will
occur in future.
- Two types - escape
and avoidance
- Escape -
learns to make a response to end the aversive situation or event (Buckle
seat belt to stop buzzing noise. Taking an aspirin to end headache.)
- Avoidance -
learns to respond to a stimulus that signals an impending unpleasant event,
avoid conditions associated with dreaded or aversive outcomes (filling
up the gas tank before you run out of gas, phobias).
Punishment
- When the behavior occurs, an
unpleasant or aversive event follows.
- Punishment decreases
the likelihood of the behavior occurring in future.
- If behavior
increases or doesn’t change, it wasn’t
punishment!
- Positive Punishment – application,
some aversive occurs to decrease behavior – spanking, child
misbehaves and gets yelled at
- Negative Punishment –
removal, something desirable is taken away to decrease behavior - being
grounded, taking away a favorite toy, if you misbehave, I’ll turn
off the TV
Punishment
Issues
- Often doesn’t extinguish the behavior – rather suppresses
behavior when punishing agent present
- Indicated behavior unacceptable, but doesn’t help develop
more effect behavior – need to reinforce appropriate behavior
- May cause negative feelings, aggression
- Most effective when applied during misbehavior or immediately
afterwards – interrupting behavior best since remove rewarding
aspects
- Must be applied consistently
Schedules of Reinforcement
- Continuous
- Reinforce every
time behavior occurs
- Improves initial
learning., most effective way to condition new responses
- Extinction occurs quickly
when reinforcement stops.
- Partial or
Intermittent
- Reinforce some,
but not all, responses.
- Better for the
maintenance of behavior.
- Resistant to
extinction (partial reinforcement extinction effect).
Partial Reinforcement: Ratio Schedules
Fixed Ratio
- Ratio: Based
on number of responses.
- Fixed: The
number of responses before reinforcement doesn’t
change.
- I’ll
pay you $30 for every car you wash.
- Behavior.:
High levels of response with pause after receiving
reinforcement. Low
resistance to extinction
· Variable Ratio
- Ratio: Based
on number of response.
- Variable: The
number of responses needed before reinforcement changes.
- Based on an average.
Slot machines pay off after an average number of lever pulls.
- Behavior:
High levels of response. Resistant to extinction.
Partial Reinforcement: Interval Schedules
Fixed Interval
- Interval:
Delivery of reinforcement dependent on the passage of time.
- A response is required.
- Fixed: The
period of time does not change.
- Picking up your
pay every Friday. Elections in the U.S.
- Behavior:
Lower levels of response. Scallop effect - increase
responding as get closer to the time you will be reinforced. Low
resistance to extinction
· Variable Interval
- Interval:
Delivery of reinforcement dependent on the passage of time.
- Variable: The
per. of time changes.
- Based on an average. Waiting
for a bus.
- Behavior:
Lower levels of steady response. High resistance to extinction
Complex Behaviors
- Shaping - Method
of successive approximation.
- Reinforcing behavior that are increasingly similar to
desired behavior.
- Chaining - Useful for
teaching a series of behavior that fit together.
- Reinforcing each
behavior as it occurs, rather than reinforcing at the end of the
sequence.
- Superstitious Behavior - Assigning
relationships between stimuli because of connections that are supposed to
exist between them. Falsely believe connection between act and consequence
- Pressing the
elevator button many times to make the elevator show up faster
Learning is not same as performance
- Latent Learning
- Rat placed in maze
without reinforcement. Shows little decrease in running times or number
of errors.
- The trial after the
first reinforcement - rat greatly improves
- Rat in group 1
reinforced for running the maze correctly, while the rat in group 2 never
reinforced. Group 3 rat reinforced only on 11th run. Rat had
learned the maze but not motivated to perform well until reinforced,
demonstrating latent learning had occurred.
Observational Learning
- Modeling
- Learning occurs by observing
others’ behaviors and the consequences of their behavior. (Vicarious
learning)
- Bandura’s
early studies
- Children watched a
film of adults hitting & kicking a doll.
- Adults were
reinforced, punished, or no consequence for their behavior.
- Children who saw the
adults reinforced were more aggressive with
the doll than the children who saw the other two versions of the film.
- Condition of adult
ignoring doll and playing with other toys. Also control group who didn’t
see a model. Ignoring doll group less aggressive than any of the other
groups (including control)
Process of Modeling
- Attention - must pay
attention to model’s behavior and the
consequences of that behavior.
- Retention - be able
to recall what was observed.
- Reproduction - motor
ability to reproduce model’s behavior.
- Motivation
want to reproduce behavior. (vicarious reinforcement)