learning Flashcards
why is it important to learn about classical and operant conditioning?
fundamental forms of learning, drug addiction relevance, well controlled experiments, AI applications, influential frameworks, types of learning are very similar across species (allows cross species comparisons), functional analogues, wider array of experimental techniques for non-human animals
what types of learning are non-declarative (implicit)?
procedural (skills, habits), priming, perceptual, classical, non-associative
what types of learning are declarative (explicit)?
facts and events (episodic and semantic)
what type of paradigms are classical and operant conditioning?
experimental paradigms (lead to highly influential frameworks for associative learning)
involves the pairing of 2 stimuli (CS + US). the US is assciated with an innate UR, this UR after learning then becomes associated with the CS to produce a CR. what type of conditioning is this?
classical
give some examples of unconditioned stimuli
water, food, hot shower, loud noise (aversive S)
give some examples of unconditioned responses
salivation, approach, licking, avoidance behaviour (for aversive S)
the CS and US can be what 2 things?
temporally segregated or overlapping
what are the typical time scales for the temporally segregated CS and US?
100ms/s
in classical conditioning, what is measured?
unconditioned responses
in classical conditioning, what is controlled by the experimenter?
unconditioned stimuli
Before: US > ? NS > ? during: US + NS > ? after: CS > ?
UR, no CR, UR, CR
common paradigms for aversive conditioning involve what types of stimuli?
eye-blink conditioning, tail pinching, electric shocks
why is aversive conditioning important to learn about?
in order to understand formation of phobias, anxiety disorders, protection mechanisms
why is aversive conditioning often single trial learning?
powerful so can be effective after first trial
give an example of aversive conditioning where longer time scales are needed
food poisoning (taste aversion testing)
classical conditioning =
operant conditioning =
stimulus-response associations, action-outcome associations
action =
outcome =
more generalised than responses in classical C, outcome = reinforcement or punishment
what are other associative learning paradigms/variations in more complex behaviours?
extinction, spontaneous recovery, generalisation, discrimination, second order conditioning, partial reinforcement, blocking
reduction shown in responding to the conditioned stimulus (CR association reduced) =
extinction
what is extinction and how is it examined?
looks at how strong/persistent the associations are that are formed during conditioning. examines what happens after conditioning when removing the US and presenting the CS alone
conditioned response comes back after a delay from conditioning =
spontaneous recovery
why would a second extinction period have a higher response strength than at the end of the first extinction period?
there is a period of delay between extinction trials that causes spontaneous recovery > response to CS returns
what happens during extinction?
CR decreases as extinction is new learning > start to form a new association opposite to the one previously formed
what does spontaneous recovery suggest?
that the original learning is not completely erased so response strength goes back up
what are the disadvantages of spontaneous recovery? (response recovery to conditioned association)
hard to get over phobias and addiction, could cause interference with new memories, extinction is specific to context so difficult to lose an association response
what are the advantages of spontaneous recovery?
flexibility to learning, adaptive (allows us to adjust)
ability to respond (CR) to the new stimulus (CS2) which resembles the original CS =
generalisation
idea of introducing another stimulus similar to original CS to form associations to a more general situation =
generalisation
why is generalisation an important learning mechanism?
avoids that we are overtrained to specific examples, in real life similar stimuli often have similar properties
ability to NOT respond (CR) to the new stimulus (CS2) which resembles the original CS =
discrimination
idea that a similar stimulus is not followed by the US as you can discriminate between 2 similar stimuli so a response association isn’t formed with the new stimuli =
discrimination
why is discrimination an important learning mechanism?
allows us to detect important differences
during training, the US is only paired with CS1 but not CS2. does this support stimulus generalisation or discrimination?
discrimination
generalisation and discrimination = ______ types of learning
opposite
need to be able to both _______ and _______ stimuli
generalise, discriminate
what do both generalisation and discrimination have interesting connections with?
cognitive abilities related to the formation of categories
after consistent pairing of CS1 with the US, CS1 can now serve as a US for further conditioning of another stimulus (CS2) =
second order conditioning
this paradigm can lead to the learning of longer chains of associated stimuli. what is it?
second order conditioning
single presentation of a CS-US sequence =
trial
consists of several trials and typically has specific parameters =
block
consists of one of more blocks that are usually separated by longer time intervals =
session
involves randomly interspersed trials where the CS is not followed by the US =
partial reinforcement
what does partial reinforcement slow down?
both acquisition (conditioning) and extinction learning
what is a practical issue with partial reinforcement?
randomisation for the design of experiments, problematic patterns can make the trial type PREDICTABLE > ppt/animal could get used to whether US is present or not
time gap in between CS and US =
trace conditioning
CS overlaps with US with no delay or time gap in between =
delay conditioning
what does trace conditioning involve if it requires the hippocampus?
neural mechanisms
is trace or delay conditioning easier to learn?
delay
how is blocking carried out?
pair CS1 with US > after some training add CS2 (CS1 + CS2 presented together and followed by US)
why with blocking, does the presentation of CS2 alone (not paired with CS1) not trigger a response?
learning is blocked as learning was already complete > CS1 precedes the US so no additional learning
idea of operant conditioning was developed by?
Thorndike (Thorndike law of effect)
responses that create a pleasant outcome = ____ likely to happen again in a similar situation
more
responses that produce a negative outcome = ____ likey to happen again in the situation
less
event that increases the likelihood of an action =
reinforcer
event that decreases the likelihood of an action =
punisher
presentation of food following a lever press would be what type of reinforcement?
+
add/increase a pleasant stimulus > behaviour is strengthened =
positive reinforcement
reduce/remove an unpleasant stimulus > behaviour is strengthened =
negative reinforcement
present/add an unpleasant stimulus > behaviour is weakened =
positive punishment
reduce/remove a pleasant stimulus > behaviour is weakened =
negative punishment
more complex realistic learning situations often involve what type of reinforcement?
partial
behaviour is reinforced after a specific number of responses = what reinforcement schedule?
fixed ratio
behaviour is reinforced after an average but unpredictable number of responses = what reinforcement schedule?
variable ratio
behaviour is reinforced for the first response after a specific amount of time has passed = what reinforcement schedule?
fixed interval
behaviour is reinforced for the first response after an average but unpredictable amount of time has passed = what reinforcement schedule?
variable interval
reinforcement schedules ______ affect the pattern and number of responses
strongly
process of guiding behaviour to the desired outcome through the use of intermediate stages in operant behaviours =
shaping
what is the advantage of shaping?
allows learning of complex sequences by dividing the learning goal into sub goals (smaller steps)
explain how training a rat to press a lever can be learnt through shaping (reinforcing individual steps)
first reinforce approaching lever > then touching lever > then pressing it