Chap 6: Learning

Ivan Pavlov

Ivan Pavlov

Found that dogs learned to pair the sounds in the environment where they were fed with the food that was given to them and begin to salivate simply upon hearing the sounds (classical conditioning)

Classical Conditioning

Learning to associate neutral stimuli (e.g. sounds) with stimuli that produces a reflexive, involuntary response (e.g. food) and will learn to respond similarly to the new stimulus as they did to the old one (e.g. salivate)

Unconditioned Stimulus (US)

A stimulus that evokes an unconditioned response without any prior conditioning (no learning needed for the response to occur)

Unconditioned Response (UR)

An unlearned reaction/response to an unconditioned stimulus that occurs without prior conditioning

Conditioned Stimulus (CS)

A previously neutral stimulus that has, through conditioning, acquired the capacity to evoke a conditioned response

Conditioned Response (CR)

A learned reaction to a conditioned stimulus that occurs because of prior conditioning

Delayed Conditioning

CS is presented before the US and it (CS) stays on until the US is presented. Works the best especially if delay is short

Trace Conditioning

The presentation of the CS, followed by a short break, followed by the presentation of the US

Simultaneous Conditioning

CS and US are presented at the same time

Backward Conditioning

US is presented first and is followed by the CS. This method is particularly ineffective.

Extinction

• The process of unlearning a behaviour (taken place when CS no longer elicits the CR).

• Achieved by repeatedly presenting the CS without the US, thus breaking the association between the two.

Spontaneous Recovery

Sometimes, after a conditioned response has been extinguished and no further training of the animals has taken place, the response briefly reappears upon presentation of the conditioned stimulus.

Generalization

Often animals conditioned to respond to a certain stimulus will also respond to similar stimuli, although the response is usually smaller in magnitude.

Discrimination

Training subjects to tell the difference between various stimuli

John B. Watson and Rosalie Rayner

Conditioned a little boy named Albert to fear a white rat.

Albert + White Rat Experiment

• Albert initially liked the white, fluffy rat, but by repeatedly pairing it with a loud noise, Watson and Rayner taught Albert to cry when he saw the rat.

• Loud noise is the US because it elicits the involuntary, natural response of fear (UR).

• Rat is a neutral stimulus that becomes the CS, and the CR is crying in response to presentation of the rat alone.

• Albert also generalized, crying in response to a white rabbit, a man’s white beard, etc.

Aversive Conditioning

The process by which a noxious or unpleasant stimulus is paired with an undesired behavior.

Second/higher-order conditioning

Once a CS elicits a CR, it is possible, briefly, to use that CS as a US in order to condition a new response to a new stimulus.

Conditioned Taste Aversions

• A learned association between the taste of a particular food and illness such that the food is considered to be the cause of the illness

• Can result in powerful avoidance responses on the basis of a single pairing

• The CS must be salient (distinctive or prominent) in order for us to avoid it

John Garcia and Robert Koelling

• Performed a famous experiment illustrating how rats more readily learned to make certain associations than others.

• Used 4 groups of subjects in their experiment and exposed each to a particular combination of CS and US

• Rats learned to associate noise with shock and unusual-tasting water with nausea

• They were unable to make the connection between noise and nausea and between unusual-tasting water and shock.

Linking loud noise → shock and unusual-tasting water → nausea seems to be adaptive

Garcia effect: the ease with which animals learn taste aversions

Edward Thorndike

Conducted a series of famous experiments using a cat in a puzzle box and created the Law of Effect

Cat in a Puzzle Box Experiment

• Hungry cat was locked in a cage next to a dish of food; cat had to get out of the cage in order to get the food

• Found that the amount of time required for the cat to get out of the box decreased over a series of trials.

• As the amount of time decreased, the cat did not seem to understand, suddenly, how to get out of the cage

• Led Thorndike to assert that the cat learned the new behaviour without mental activity but rather simply connected a stimulus and a response

Law of Effect

• If the consequences of a behaviour are pleasant, the stimulus-response (S-R) connection will be strengthened and the likelihood of the behaviour will decrease.

• If the consequences of a behaviour are unpleasant, the S-R connection will weaken and the likelihood of the behaviour will decrease.

• Thorndike used the term instrumental learning to describe his work because he believed the consequence was instrumental in shaping future behaviours

B. F. Skinner

Used a Skinner box to research animal learning (has a lever/disk to press in order to get food)

Reinforcement

• Anything that makes a behaviour more likely through reward or removing something unpleasant

Positive Reinforcement

Addition of something pleasant

Negative Reinforcement

Removal of something unpleasant

Escape Learning

Allows one to terminate an aversive stimulus

Avoidance Learning

Enables one to avoid the unpleasant stimulus altogether

Punishment

Anything that makes a behaviour less likely through unpleasant consequences

Positive Punishment

Addition of something unpleasant

Negative Punishment (Omission Training)

Removal of something pleasant

Shaping

The production of new forms of operant behavior by reinforcement of successive approximations to the behavior

Chaining

Teaching method based on task analysis, wherein all the smaller units of behavior comprising a complex skill or task is identified and broken down first and the series of related behaviors is taught in a step by step manner.

Primary Reinforcers

Reinforcers that are naturally rewarding. (e.g. food, water, rest, etc.)

Secondary Reinforcers

Things we have learned to value. (e.g. praise, chance to play a video game, etc.)

Generalized Reinforcer

Money is a special kind of secondary reinforcer, a generalized reinforcer, because they can be traded for virtually anything

Token Economy

In a token economy, every time people perform a desired behaviour, they are given a token (e.g. money). Periodically, they are allowed to trade their tokens for any one of a variety of reinforcers.

Premack Principle

Whichever of 2 activities is preferred can be used to reinforce the activity that is not preferred.

Continuous Reinforcement

Rewarding the behaviour each time (best when first teaching it)

Partial-Reinforcement Effect

Increased resistance to extinction after intermittent reinforcement rather than after continuous reinforcement.

Fixed-Ratio Reinforcement Schedule

Reinforcement is delivered after a set number of responses.

Fixed-Interval Reinforcement Schedule

Reinforcement is delivered after a behaviour is performed following the passage of a fixed amount of time.

Variable-Ratio Reinforcement Schedule

Reinforcement is delivered after a variable number of responses.

Variable-Interval Reinforcement Schedule

Reinforcement is delivered after a behaviour is performed following the passage of a variable amount of time.

Contiguity model (Pavlovian model)

• The more times 2 things are paired, the greater the learning that will take place

• Contiguity (togetherness) determines the strength of the response

Robert Rescorla

Revised the contiguity model to take into account a more complex set of circumstances; created the contingency model

Contingency Model

• A is contingent upon B when A depends upon B and vice versa.

• The predictability of occurence of one stimulus from the presence of another

Contiguity vs Contingency Models

Pavlov’s contiguity model of classical conditioning holds that the strength of an association between two events is closely linked to the number of times they have been paired in time.

Rescorla’s contingency model of classical conditioning reflects more of a cognitive spin, positioning that it is necessary for one event to reliably predict another for a strong association between the two to result.

Observational Learning

Process of learning by watching the behaviours of others

Modeling

• Two components: observation and imitation

• Mental representation of the observed behaviour must exist in order to enable the person or animal to imitate it

Albert Bandura

Studied modeling in the Bobo Doll Experiment

Bobo Doll Experiment

Demonstrated that children who had watched an adult being violent with a Bobo doll were more likely to behave aggressively toward the doll than were children who had watched an adult being nonviolent toward it.

Latent Learning

• Latent means hidden, and latent learning is learning that becomes obvious only once a reinforcement is given for demonstrating it.

• Behaviourists asserted that learning is evidenced by gradual changes in behaviour, but Tolman conducted a famous experiment illustrating that sometimes learning occurs but is not immediately evidenced.

Edward Tolman’s Latent Learning Experiments

• Had 3 groups of rats run through a maze on a series of trials.

→ Group 1: Got a reward each time it completed the maze; performance improved steadily over the trials

→ Group 2: Never got the reward; performance improved only slightly over the trials

→ Group 3: Not rewarded in first half, but was rewarded in second; performce improved dramatically and suddenly once it began to be rewarded

• Reasoned that these rats must have learned their way around the maze during the first set of trials; performance did not improve because they had no reason to run the maze quickly.

• Latent learning is demonstrated → The rats made a mental representation (cognitive map) of the maze during the first half

Abstract Learning

Acquiring knowledge of general or intangible material, such as the meanings of concepts and propositions and the logical and systematic relations between them

Insight Learning

Occurs when one suddenly realizes how to solve a problem.

Wolfgang Köhler’s Insight Learning Experiments

• Argued that learning oten happens in this sudden way due to insight rather than because of the gradual strengthening of he S-R connection

• Suspended a banana from a cieling wall and had several boxes, none of which was high enough to enable the chimpanzees to reach the banana. Köhler found that the chimps spent most of their time unproductively rather than slowly working towards a solution until, all of a sudden, they would pile the boxes on top of each other, climb up, and grab the banana.

• Köhler believed that the solution could not occur until the chimpanzees had a cognitive insight about how to solve the problem