Please enable JavaScript.
Coggle requires JavaScript to display documents.
Intro. to Learning Theory - Coggle Diagram
Intro. to Learning Theory
Classical conditioning
process of learning to assosciate stimuli and consequently associate events with one another
Pavlov - the idea u can condition responses by pairing stimuli so that eventually that stimuli alone elicits conditioned responsed
utilising physiological reflexes - ie salivation not learned
experiment with dogs
Food (UCS) --> salivation (UCR)
Food (UCS) + Tone (NS) --> salivation (UCR)
tone is neutral since doesn't naturally elicit the UCR
paired again and again before 3.
Tone (CS) --> salivation (CR
second-order conditioning = assosciating new stimulus with a conditioned stimulus
applications
electric fences - electric shock = UCS which produces pain - UCS - this shock is assosciated with edge of field
Little Albert - conditoning fear response
paired loud noise of banging iron bar with response (fear) with NS = white rat
stimulus generalisation
= fear response extended to other similar objects
-ves:
unethical now
subjective = no objective measure of bhvr - was Albert rll scared of the rat?
poor scientific records
led to systematic desensitisation phobia studies
The Process
Intial period of learning =
acquisition
= NS begins producing CR
importance of timing - needs to be short time btwn CS and UCs for assosciation to be made
extinction
= decrease in conditioned response when UCS no longer presented with CS = response gets weaker and weaker until none
spontaneous recovery
= if CS stopped for a while but then used again - the response can be suddently remembered
Operant Conditioning
learning from +ve or -ve consequences
influenced by Thorndike's
law of effect
ie cats learning to escape by trial and error from a box
bhvrs of +ve outcomes that satisgy organism more likely repeated, bhvrs followed by unpleasant consequences less likely to be repeated
profitless acts sampted out, rewards stamped in
Skinner's box - level pressed for food rward, recorder of no. of responses, electric shock
Reinforcement =
increases behaviour
, used to motivate bhvr occuring in future
positive: presenting reinforcing stimulus after desired bhvr, making it more likely to happen
negative = removing aversive stimulus (ie stopping electric shock) after desired bhvr to increase it
reinforcement schedules - how you time and space reinforcements makes more/less likely to occur
partial reinforcement = don't reward every time
fixed interval = reinforcement at predictable time intervals
variable interval = delibered at unpredictable time intervals
fixed ratio = delivered after predicable no. of responses
variable ratio = delivered after unpredictable no. of responses = high and stead response rate!
punishment =
decreases behaviour
positive = presenting aversive consequence after an UNdesired bhvr making less likely to happen
negative = reinforcing stimulus removed after UNdesired bhvr exhibited
positive reinforcement = more useful, more enduring bhvrs, quicker vs negative stimuli = least useful, rejected usefuleness of punishment
shaping
= reward successive approximiations of a target bhvr - reinforcing responses that resembles desired bhvr until closer and closer to desired
primary vs secondary reinforcers
: primary = natural UCR to stimulus, secondary = learned/conditioned ie the "good boy" alongside the food (primary)
primary = more primitive brain regions, secondary = newer regions like frontal cortex
linked to
token economies
= rewarding ppl for +ve bhvrs with tokens/stars (secondary reinforces) that are exhcanged for desirable things ie sweets (primary reinforcers)
Rescorla Wagner Model
delta V = a b (lamda - sum of V)
contiguity
= time btwn stimulus and outcome
expectation on any given trial is based on predictive value of all stimuli present - depends on what have already learnt
amount of learning depends on the amount of surprise (diff btwn what actually happens - lamda - and what you expect - sum of V)
delta V = change in strength of association btwn CS and US
alpha = salience of CS
beta = motivational value of US, outcome
lamda = outcome/presence of US - 0-1
sum of V - sum of assosciative strengths of stimuli
forms asymptotic curve - rate of change fast to start then slows down
blocking
= novel stimulus presented with well-established CS - predictive value doesn't change for either since no surprise = no learning - since CS predicts US
unblocking
= novel stimulus introduced and US is larger = there is surprise
acquisition curve goes up quickly when CS tells u little so surprised by UCS - learn a lot for pairing - when greater prior learning about CS learn little
BUT fails to account for spontaneous recovery, rapid reacquisiton, latent inhibition
Observational learning
latent learning
= occurs without obvious reinforcement
children learning by watching actions of parents but only demonstrate at later date
Tolman's rats - rats can learn even without immediate reinforcement, unreinforced rats developed cognitive map of maze layout - able to find way through maze just as quickly as comparison grp rewarded with food
learning by watching, imitating, modelling - using mirror neurons
finding that children as young as 14 mths imitate type of play on video from observing that tape of play on TV
Bandura
Bobo Doll experiment
(study 2) vicarious reinforcement = watching models' aggression being rewarded or punished leads to imitated more or less aggressive actions of model - reinforcement gained by watching another person's
(study 1) - modelling of aggressive/quiet and subdued/ or no teacher - children told experimenter's best toys that decided to reserve - next room contained some aggressive/non-aggressive toys - children who observed aggressive model made more imitative aggressive responses - even picked up NOVEL language/guns (not modelled -
emulation
)
links to violent TV shows - Fitzpatrick et al (2016) - even cues in films can trigger aggressive acts, tho effects are inconclusiv, -ve effects on self-regulation
complex links btwn aggression/violent media/emotional reg/high order exec function
SLT: attention >> retention >> motor reproduction >> motvation - mediating processes occur btwn stimuli and responses
interaction btwn environmental factors and cognitive
exposure not enough, observer must pay attention 2. bhvr must be remembered to be imitated 3. must be able to physically reproduce 4. observer considering rewards and punishments to follow bhvr
Reflexes and instincts
innate behaviours = born with, not learnt/taughts
reflexes
= motor response to external stimulus, not consciously ctrlled
instinct
= innate bhvrs in response to more complex scenarios or stimuli ie migration, usually more than one action
both advantageous, but neither learned vs learned bhvrs - require experience
assosciative learning = making connections btwn event and behaviour - can study through focusing on what can be seen and measured - eg's = classical + operant conditioning
habituation
= we learn NOT to respond to stimulus that is presented repeatedly without change = become accustomed to it so eventually don't notice
used in advertising - ie assosciated celebriities with product (CS) where UCR -> CR is positive feelings for the celebrity