lol Flashcards
what is intelligence?
what you do when you dont know what to do - piaget
intelligence theories?
IQ, Catell (fluid vs. crystallized), specific vs. general, instrumental vs. social affective
aristotle?
an empiricist, formed first theory of associationism, argued that memory is dependent on associations between pairs of events - so recalling one member of the pair elicits anticipation of the other!
aristotle vs plato!
draw it out idiot
aristotles questions?
- how are associations made
- emergence of nature/nurture
- questions abt. learning in humans/animals
- are learning and memory universal?
classical vs. instrumental condtioning!
draw it out idiot
3 major dopaminergic pathways in brain (main one?)
mesolimbic
instrumental conditioning for thorndike
law of effect
- it is about S-R learning
- association strengthened if S followed by satisfying event
- association weakened if S followed by annoying event
importance of outcomes (all four)
- reward: produces good outcome
- punishment: produces bad outcome
- omission: prevents good outcome
- avoidance/escape: prevents bad outcome
importance of outcomes chart!
draw it out idiot
+ reinforcement
response produces appetitive S, reinforcement
+ punishment
response produces aversive S, punishment
(- reinforcement
response eliminates occurence of aversive S, reinforcement (escape)
(- punishment
response eliminates occurence of appetitive S, punishment (omission)
sign tracking chart
draw it out idiot
pavlovian classical example
bell (CS) —- Food (US)
\ \
\ \
\ \
salivate (CR) salivate (UR)
what is generalization?
when one CS generalizes to another!
two historic views of classical conditoning?
SR and SS
SR learning (Drawn out)
bell (CS) Food (US)
\
\
\
salivate (CR) salivate (UR)
SS learning (Drawn out)
bell (CS) ——–> Food (US)
salivate (CR) salivate (UR)
temporary separation of CS/US in classical conditioning leads to
decrease in sign tracking, increase in goal tracking
what type of tracking is resistant to outcome devaluation
sign tracking!
is it possible to respond to a CS that was never paired with the US (mechanism)?
US devaluation
- higher order conditioning
- sensory preconditoning
dog bite scenario!
draw out idiot
generalization
tendency to respond the way way to different but similar stimuli
higher (second order conditoning)
when CS can elicit CR without being paired with US
- additional NS presented after initial NS has been paired with US to elicit CR
sensory preconditioning
initially paired two NS, pairing one with US. if the other can elicit a response than sensory preconditoning has occurred
emotional classical conditioning: fear in rats
conditioned response is often freezing. The CS supresses bar pressing»_space; response is supressed. suppression is the DV of conditoned fear
supression ratio
( responding CS) / (responding CS + responding pre-CS)
what type of tracking is autoshaping in pigeons?
sign tracking!
strength of conditioning (4)!
- timing
- novelty of CS/US
- intensity of CS/US
- pseudoconditoning
four types of CS/US associations
delay
trace
simultaneous
backwards
best timing for conditoning
short ISI’s (interstimulus intervals), long ITI’s (intertrial intervals)
two types of counterfeit conditoning
sensitization
pseudoconditioning
how to win CS competetion?
- be more noticeable – overshadowing
- be there sooner – blocking
opposite of overshadowing?
compound potentation
what is compound potentiation?
when there is more conditoning to a weak stimulus if combined with a salient stimulus
difference between overshadowing and blockign
overshadowing occurs only in one phase
learning only occurs when…
CS provides new info on US!
rescorla/wagner (key!)
US modulation, blocking, surprise
mackintosh (key!)
CS modulation, attention, latent inhibtion
pearce-hall (key!)
US nd CS modulation, attnetion, negative transfer
Wagner (SOP and AESOP) (key!)
priming, affective dimension
rescor wagner three types
- novel CS – US: US is unexpected, positive prediction error, larger error = greater learning
- well learned CS – US: expected, no error prediction, no learning!
- CS-US association – no US: absence is unexpected, negative prediction, decreased association
what did rescorla/wagner contribute habituation to?
decrease in stimulus suprise with self and retrieval priming
RW assumption
CSs compete for associative weight - surprise dimension!
mathematical dimension (Draw it first)
delta V = change in associative strength
alpha = CS saliency
beta = US saliency
lambda = ceiling (asymptote of learning
V = learning parameter (how well CS predicts US)
[ lambda - V ] = surprisingness of US
when does V decrease?
with extinction!
what will strong US impact
associative strength or lambda
what will salient CS impact
rate of learning!
suppression ratio?
less learning with high suppression ratio. if new stimulus is presented with another that was previously conditioned, little or no conditioning will occur (high suppression ratio)
trick to get very efficient extinction
combine CS with other excitators!
comparator theory
conditioning to context is fundamental to contingency
(CS US and context are associated)
comparator theory doesnt explain
extinction of inhibition, latent inhibition, issues with blocking
low suppression ratio =
more conditioning
what does comparator theory suggest
we learn to ignore redundant predictors of US, abt the ineffectivenss of CS
mackintosh model what is key
alpha (CS saliency) is key!
pearce-hall?
pairing a CS with a weak US slows down condiioning when CS is paired with a strong US – negative transfer
pearce hall important!
organisms attent ot uncertain predictors! surprise!
kaye/pearce
suprisingness of US matter, alpha is high, exlpains negative transfer and latent inhibition
hybrid attnetional models (combo)
- cues with high predictive value (mackintosh)
- cues high in uncertainty (pearce-hall)
+ RW: surprisingness of US+CS!!!
primed stimulus
= less suprising
two types of priming
- self-generated (Via sensory memory)
- retrieval-generated (from long term memory via. retrieval)
what kind of priming is blocking?
retrieval generated!
SOP
originally PDP, standard operating procedure (sometimes oppenent process), address time, connectionist model
both types of priming can…
reduce surprise
what does SOP suggest
- processing is parallel not serial
- processing is distributed not localized
AESOP
affective extention of SOP, adds emotional qualities
CS gets linked to sensory and affective nodes!
contiguity?
CS occurs with US
contingency?
CS predicts US
classical vs. AESOP view of dog bite!
draw out idiot
pearce said AESOP is…
configural - the whole set is conditioned to US
what does adding diff CSs do?
weaken the value, eg. only AB loses 33% of conditioning
classical vs. operant
operant outcome is contingent on response. both have learning curve, both have extinction.
who believes in contiguity and what is at its core
guthrie, stimulus elemtns
thorndike
guthrie
tolman
thorndike: S-R, rewards enforce
guthrie: S-R, rewards are very salient
tolman: S-S, rewards motivate
two main methods of skinner box
1.discrete trial: many stand alone trials, controlled by experimenter
2.free-operant: animal controls apparatus
EDT advantages and disadvantages
- adv. less mistakes, no negative emotions, very fast training
- Dadv. modifications/reversals in training are difficult
conditoned/secondary reinforcers are
like clicker. classically conditioned discriminative stimulus.
clicker = pavlovian for food
negative contrast?
human and animals both sensitive to reinforcers of different vlues
why use clicker (3)?
- necessary bc of distance
- spatital/temporal precision is necessary
- training behaviour chains
types of schedules
fixed ratio
variable ratio
fixed interval
variable interval
best schedules?
variable ratio, variable interval
other types of schedules?
compound
1. mixed - alternating 2
2. multiple - alternating 2 types, w diff DS
3. tandem - one then next
4. chained - tandem/multiple + diff DS
what is matching law (when organism chooses)
rate of responding on a particular choice will match the relative role of reinforcement on that choice
- eg. pigeon pecks twice as much at lever every 1 min compared to 2min
what explains matching law?
- animals trying to maximize reinforcement
- melioration = improve local rate of reinforcement
drive reduction
hull, homeostatic, deficit, incentives
premack
differential probability - contingency
- more probable responses reinforce less probable responses
behavioural regulation
bliss point, about max reinforcement
selection by consequences
learning is like natural selection – only an analogy
- needs trait and selective gent
selection by consequences requires
- variation
- fitness consequences
- mode of inheritance
- limited resources
motivation
hypothetical constuct, needs intervening variables
do you need reinforcement for learning
no
learning will always extinguish if not reinforced
no
habit learning =
instrumental in nature, motor learning, basal ganglia
motor learning, motivation and emotion =
all linked in brain, heavily driven by dopaminergic system
drive theory - resistance to extinction depedns on
- drive (degree of food deprivation)
- habit (amount of prior reinforcement)
drive theory formul
behaviour strength = D x H
incentive theory formul
behaviour strength = D x H x K
or
behaviour strength = need x learning x incentive
incentive =
motivational effect of reward (this idea comes from classical conditoning!)
motivational states =
occasion settors, facilitators
OS are…
environmental cues (CSs) that faciliate conditoning. CS that confirms a CS-US pairing
elation vs depression
elation = positive contrast effect
depression = negative contrast effect
over-learning extinction effect
many rewarded trials can increase the rate of extinction
sequential theory variables defined
N = nonreinforced trials
r = reinforced trials w small rewards
R = reinforced trials w big rewards
sequential theory most resistant to extinction?
rNR!
