2009. How Do We Learn? A relatively permanent change in behavior that results from experience

10/20/2009 Learning and Cognitive Processes Understanding Psychology Chapter 2 How Do We Learn? We learn by association. Our minds  naturally connec...
Author: Clara Hicks
6 downloads 2 Views 839KB Size
10/20/2009

Learning and Cognitive Processes Understanding Psychology Chapter 2

How Do We Learn? We learn by association. Our minds  naturally connect events that occur in  sequence.  q

learning A relatively permanent change in behavior that results from experience

4

Associative Learning

How Do We Learn? 2000 years ago, Aristotle  suggested this law of  association. 

Learning to associate one stimulus with another.

Then 200 years ago Locke and  Hume reiterated this law.

5

6

1

10/20/2009

Associative Learning

Associative Learning

Learning to associate one stimulus with another.

Learning to associate a response with a consequence.

8

7

Three Basic Types of Learning:

Associative Learning Learning to associate a response with a consequence.

• Classical conditioning • Operant conditioning • Modeling

9

Classical Conditioning

Classical Conditioning

Sovfoto

Ideas of classical conditioning originate from old  philosophical theories.  However, it was the  Russian physiologist Ivan Pavlov who elucidated  classical conditioning. His work provided a basis  f l for later behaviorists like John Watson. b h i i lik J h W

Ivan Pavlov (1849‐1936)

12

2

10/20/2009

Pavlov’s Experiments

Pavlov’s Experiments

Before conditioning, food (Unconditioned  Stimulus, US) produces salivation  (Unconditioned Response, UR). However, the  tone (neutral stimulus) does not.

During conditioning, the neutral stimulus (tone)  and the US (food) are paired, resulting in  salivation (UR). After conditioning, the neutral  stimulus (now Conditioned Stimulus, CS) elicits  salivation (now Conditioned Response, CR) li i ( C di i dR CR)

13

14

Acquisition Acquisition is the initial learning stage in  classical conditioning in which an association  between a neutral stimulus and an  unconditioned stimulus takes place.

classical conditioning

1. In most cases, for conditioning to occur, the  neutral stimulus needs to come before the  unconditioned stimulus. 2. The time in between the two stimuli should  be about half a second.

Controlling an animal’s or a person’s responses in such a way that an old response becomes attached to a new stimulus 15

Classical Conditioning • Occurs gradually • Each pairing of the CS with the UCS strengthens the conditioning • Timing Ti i is i an issue. i Conditioning C diti i is i strongest when the CS is introduced just before the UCS

+

3

10/20/2009

Stimulus Generalization

Stimulus Discrimination Discrimination is the learned ability to distinguish  between a conditioned stimulus and other stimuli that  do not signal an unconditioned stimulus.

Tendency to respond to  stimuli similar to the CS is  called generalization. Pavlov  conditioned the dog’s  g salivation (CR) by using  miniature vibrators (CS) on  the thigh. When he  subsequently stimulated  other parts of the dog’s  body, salivation dropped. 19

20

Extinction

Spontaneous Recovery

When the US (food) does not follow the CS  (tone), CR (salivation) begins to decrease and  eventually causes extinction.

After a rest period, an extinguished CR (salivation)  spontaneously recovers, but if the CS (tone) persists  alone, the CR becomes extinct again.

21

Classical Conditioning & Human Behavior • O. Hobart & Mollie Mowrer used classical conditioning to find a practical solution to bedwetting. • Bell and pad device – two metallic sheets perforated with small holes and attached by wires to a battery-powered alarm.

22

Classical Conditioning & Human Behavior • Placed under the child’s bedsheets. When the sleeping child begins to moisten the sheet, the first drops of urine complete the circuit and cause the alarm to go off, off waking the child so that he or she can use the bathroom. • After several pairings of the full bladder (CS) and the alarm (UCS), the child is able to awaken to the sensation of a full bladder without the help of the alarm.

4

10/20/2009

Taste Aversions

Taste Aversions • John Garcia and R.A. Koelling (1966) tested rats by associating flavored water with flashing lights and an electric shock. • Gustavson, Gustavson et al (1974) taught coyotes to hate the taste of sheep by giving them shots after eating sheep that made them sick. Teaching the coyotes to avoid eating sheep satisfied sheep ranchers and naturalists.

• The person tries out a new food. • Several later, S l hours h l t he h or she h gets t violently sick. • He or she will probably blame the new food and in the future will become nauseated by even seeing the new food, even if it is not the cause of the illness.

Applications of Classical  Conditioning

The Case of Little Albert • John B. Watson and Rosalie Rayner experimented on a well-adjusted 9 monthold human infant with classical conditioning. conditioning

John B. Watson

Brown Brothers

Watson used classical  conditioning procedures to  p g develop advertising  campaigns for a number of  organizations, including  Maxwell House, making the  “coffee break” an American  custom.

27

The Case of Little Albert • Albert showed no fear when he was presented with a variety of objects including a rat, blocks, a rabbit, a dog, a monkey, masks with and without hair, cotton, wool and b i newspapers. All stimuli burning ti li were neutral t l to him.

The Case of Little Albert • When 11 months old, they placed a white furry rat in front of Albert. Every time he reached for the rat, they would it a metal bar with a hammer behind him him, making a loud noise. The noise paired with the sight of the rat elicited fear. After a few pairings, Albert feared not only rats, but anything that was white and furry.

5

10/20/2009

Criticisms of Little Albert Experiments • Made an otherwise healthy child fearful. • The researchers made no attempt to extinguish the CR even though they knew he was leaving the experiment. experiment • Mary Cover Jones, one of Watson’s students, developed methods for reducing people’s fears by pairing things which people are fearful of with pleasant stimuli.

Ethical Principles • Planning research – Ethical problems are resolved before the research starts • Responsibility – Psychologists are responsible for the dignity and welfare of participants • Compliance with law and standards – Follow all federal and state laws, as well as professional guidelines

Ethics of Research with Human Participants • Ethical procedures have been developed by the American Psychological Association in research with human participants

Ethical Principles • Research responsibilites – except for anonymous surveys, nauralistic observations and similar research, psychologists reach an agreement regarding rights and responsibilities of both subjects and researchers before research is begun. • Informed consent – If consent is required, psychologists obtain a signed informed consent before starting any research with a subject. They inform subjects of the nature of the research so that they are free to take part or decline to take part.

Ethical Principles

Reinforcement

• Deception in research – Deception is only used in research if no better alternative is available. Under no condition may decption be used about any negative aspects that might influence a subject’s willingness to participate

• B.F. Skinner has been the psychologist most often associated with operant conditioning. • He believed that most behavior is influenced by one’s history of rewards and punishments • Reinforcement is defined as a stimulus or event that affects the likelihood that an immediately preceding behavior will be repeated.

6

Skinner’s Experiments

Operant Chamber

Skinner’s experiments extend Thorndike’s  thinking, especially his law of effect. This law states that rewarded behavior is likely to occur  again.

From The Essentials of Co onditioning and Learning, 3rd Edition by Michael P. Domjjan, 2005. Used with permission by Thomson Learning, Wa adsworth Division

10/20/2009

Using Thorndikeʹs law of effect as a starting  point, Skinner developed the Operant chamber,  or the Skinner box, to study operant  conditioning. Walter Dawn/ Photo Researc

Operant Chamber

Yale University Library

37

38

Operant vs. Classical Conditioning • In classical conditioning, the experimenter presents the CS and UCS independent of the subject’s behavior. Reactions to the CS are then observed. • In operant conditioning, the subject must engage in a behavior in order for the programmed outcome to occur. • Operant conditioning is learning from the consequences of behavior

The operant chamber, or Skinner box, comes  with a bar or key that  an animal manipulates  p to obtain a reinforcer  like food or water. The  bar or key is connected  to devices that record  the animal’s response. 39

Types of Reinforcers

Primary & Secondary Reinforcers

Reinforcement: Any event that strengthens the  behavior it follows. A heat lamp positively  reinforces a meerkat’s behavior in the cold.

1. Primary Reinforcer: An innately reinforcing  stimulus like food or drink. Co ditio ed ei o ce : A A learned  ea ed 2.. Conditioned Reinforcer: reinforcer that gets its reinforcing power  through association with the primary  reinforcer.

Reuters/ Corb

41

42

7

10/20/2009

Reinforcement Schedules

Immediate & Delayed Reinforcers 1. Immediate Reinforcer: A reinforcer that  occurs instantly after a behavior. A rat gets  a food pellet for a bar press.

1. Continuous Reinforcement: Reinforces the  desired response each time it occurs. a tia ei o ce e t: Reinforces a  ei o ces a 2.. Partial Reinforcement: response only part of the time. Though this   results in slower acquisition in the  beginning, it shows greater resistance to  extinction later on. 

2 D 2. Delayed Reinforcer: l dR i f A i f A reinforcer that is  th t i delayed in time for a certain behavior. A  paycheck that comes at the end of a week. We may be inclined to engage in small immediate  reinforcers (watching TV) rather than large delayed  reinforcers (getting an A in a course) which require  consistent study.

43

44

Interval Schedules

Ratio Schedules 1. Fixed‐ratio schedule: Reinforces a response  only after a specified number of responses.  e.g., piecework pay.

1. Fixed‐interval schedule: Reinforces a  response only after a specified time has  elapsed. (e.g., preparing for an exam  only when the exam draws close.)

2 V 2. Variable‐ratio schedule: i bl ti h d l Reinforces a  R i f response after an unpredictable number of  responses. This is hard to extinguish because  of the unpredictability. (e.g., behaviors like  gambling, fishing.)

2. Variable‐interval schedule: Reinforces a  response at unpredictable time  intervals, which produces slow, steady  responses. (e.g., pop quiz.) 45

Schedules of Reinforcement

46

Stimulus Control • In operant conditioning, stimuli associated with receiving rewards or punishments become signals for particular behaviors. • Organisms learn to generalize among and discriminate between stimuli that serve as signals in operant conditioning.

47

8

10/20/2009

Stimulus Control • In such cases, the signal is called a secondary reinforcer or conditioned reinforcer.

Stimulus Control • In such cases, the signal is called a secondary reinforcer or conditioned reinforcer.

– Without the conditioning process process, it would be a neutral stimulus, with no intrinsic value – With conditioning, almost any stimulus can acquire almost any value.

– Without the conditioning process process, it would be a neutral stimulus, with no intrinsic value – With conditioning, almost any stimulus can acquire almost any value.

Stimulus Control

Stimulus Control

• Wolfe (1936) did a study with chimpanzees in which poker chips which have no value for chimps (not edible or fun to play with) developed value to the chimps, much as money has value to humans. • A machine dispensed peanuts or bananas, when chimps inserted a poker chip. The chimps “earned” poker chips by pulling down on a heavily weighted bar.

• With repetition, poker chips became reinforcers themselves. Chimps would work for them, save them and sometimes try to steal them from one another. • Another example, smiles for newborn babies. No value of their own to newborns, but become signals that a baby will be picked up, cuddled, perhaps fed. The smiles become a reward of their own in time.

Aversive Control

Aversive Control

• Reinforcement is often used to refer only to pleasant consequences of behavior • But psychologists use the term to refer to y g that increases the frequency q y of anything an immediately preceding behavior • Aversive or unpleasant consequences influence much of our daily behavior • We refer to conditioning which uses this type of reinforcer as aversive control.

• There are two ways in which aversive stimuli can affect our behavior: – As negative reinforcers – As punishers p

• Negative reinforcement – painful or unpleasant stimuli is removed or not applied at all. Two types: – Escape conditioning – Avoidance conditioning

9

10/20/2009

Punishment

Punishment

An aversive event that decreases the behavior it  follows. 

Although there may be some justification for  occasional punishment (Larzelaere & Baumrind,  2002), it usually leads to negative effects. 1. 2. 3. 4.

55

Results in unwanted fears. Results in unwanted fears. Conveys no information to the organism. Justifies pain to others. Causes unwanted behaviors to reappear in its  absence. 5. Causes aggression towards the agent. 6. Causes one unwanted behavior to appear in  place of another.

56

Escape Conditioning

Avoidance Conditioning

• Escape conditioning – behavior allows the subject to escape the aversive stimulus already presented.

• Avoidance conditioning – subject’s behavior prevents an unpleasant situation from happening

– Child served a food it doesn’t doesn t like – Whines and gags – The disliked food is removed – Whining and gagging are reinforced and will probably be used again when another food it dislikes is introduced

– If child child’s s whining and gagging kept the mother from even serving the disliked food, the behavior would be classified as avoidance conditioning

Cognition & Operant Conditioning

Latent Learning

Evidence of cognitive processes during operant  learning comes from rats during a maze  exploration in which they navigate the maze  without an obvious reward. Rats seem to  develop cognitive maps, or mental develop cognitive maps, or mental  representations, of the layout of the maze  (environment).

Such cognitive maps are based on latent  learning, which becomes apparent only when  an incentive is given (Tolman & Honzik, 1930).

59

60

10

10/20/2009

Punishment

Intrinsic Motivation Intrinsic Motivation: The desire to perform a  behavior for its own  sake.

• In punishment, an unpleasant consequence occurs and decreases the frequency of the behavior that produced it. • Negative reinforcement and punishment act in opposite ways:

Extrinsic Motivation: The desire to perform a  behavior due to  promised rewards or  threats of punishments. 61

Disadvantages of Using Aversive Control • Aversive stimuli can produce unwanted side effects such as rage, aggression, and fear • Then instead of one problem behavior to change, h th there may b be additional dditi l behaviors b h i such as aggressive behavior toward other children. • A second problem is that people learn to avoid the person delivering the aversive stimuli. Therefore parents and teacher s have less opportunity to correct a child’s inappropriate behavior.

– In negative reinforcement, escape or avoidance behavior is repeated, and increases in frequency. – In punishment, behavior that is punished decreases or is not repeated

Factors That Affect Learning • Feedback • Transfer • Practice P ti

Feedback

Transfer

• Finding out the results of an action or performance • Without feedback, may repeat same errors many times • Without feedback, would not be reinforced for correct actions to encourage continuing

• Transfer of a skill that you have already learned to help you learn a new skill. • Ex.: If you have already learned to play the p , yyou can use those skills to saxophone, help you learn to play the clarinet • Positive transfer – previously learned responses help you master a new task • Negative transfer – previously learned responses actually make it harder to learn the new task

11

10/20/2009

Practice • The repetition of a task • Helps to bind responses together • Practice works best when done over a period i d off ti time iinstead t d off allll att once. • Mental practice can be used by athletes. Though not as effective as physical practice, better than not practicing.

Harry Harlow • Harry Harlow (1949) performed experiments to show that animals could learn to learn. y had to find which lid raisins were • Monkey hid under. First, using different sizes and shapes of wooden lids, color was the key, always under green lids. • When solved, changed the experiment to shape of lid. • Eventually monkey could solve the problem without making any more than i t k

Hiroto (1974) • One group of college students was able to turn off an unpleasant loud noise • Another group was not able to turn off the noise • Later placed in a situation where all they had to do to turn off the loud noise was to pull a lever, the group who could turn it off before turned it off, the other group didn’t even try

Learning Strategies • Problem-solving skills may transfer from one type of problem to other similar problems • If a strategy works works, a person person, or an animal is likely to use it again

Learned Helplessness & Laziness • If a person has numerous experiences in which his actions have no effect on the world, he may learn a general strategy of learned helplessness or laziness • If rewards come without effort, a person never learns to work (learned laziness) • If pain comes no matter how hard one tries, a person gives up (learned helplessness).

Seligman’s 3 Elements of Learned Helplessness • Stability • Globality • Internality I t lit

12

10/20/2009

Stability

Globality • A person decides that the state of helplessness comes from a global reason rather than specific.

• A person’s belief that the state of helplessness results from a permanent characteristic instead of something g temporary in nature.

– Ex.: Ex : Rather than “II am not good at taking math tests” (specific), “I am just dumb” (global)

– Ex.: “I am not any good at taking tests. I always do poorly and I always will” (stable), rather than “I did poorly on the math test because I was sick or I didn’t get enough sleep last night.” (temporary)

Internality

Shaping

• A person attributes an undesirable outcome to their own inadequacies instead of to external reasons. Ex : Using a stability or a globality Ex.: focus to explain poor performance on the math test instead of an external reason: “This was a poorly designed math test.”

• A technique of operant conditioning in which the desired behavior is “molded” by first rewarding any act similar to that behavior and then requiring closer and closer approximations to desired behavior before giving the reward.

Teaching a Rat to Raise a Miniature Flag

Shaping Shaping is the operant conditioning procedure  in which reinforcers guide behavior towards the  desired target behavior through successive  approximations. Fred Bavendam/ Peter Arno

Khamis Ramadhan/ Panapress/ G

A rat shaped to sniff mines. A manatee shaped to discriminate objects of different shapes, colors and sizes.

• The experimenter places a rat on a table with a miniature flag pole in the center. • When the rat sniffs the flagpole, he gets rewarded. rewarded • He starts sniffing the flagpole a lot, hoping to get rewarded again. But the reward is withheld until the rat lifts a paw.

77

13

10/20/2009

Teaching a Rat to Raise a Miniature Flag • The experimenter continues to reward the rat as it comes closer and closer to the desired activity. • Eventually the rat raises up on its hind legs and nibbles on the cord. He is rewarded. • Suddenly he nibble on the cord and yanks it. He is rewarded and the rat begins pulling rapidly on the cord and the new response is shaped.

The Military Shapes Complicated Behaviors in Animals • During WWI, the British trained seagulls to detect enemy submarines in the English Channel. • During WWII, B.F. Skinner was involved in a project teaching pigeons to guide missiles to enemy targets. • The U.S. Navy has trained dolphins to detect enemy divers and locate undersea mines. • Sea lions have been taught to recover antisubmarine rockets.

Combining Responses: Chaining

Combining Responses: Chaining

• To learn a new skill, a person must be able to put various new responses together. • Responses that follow one another in a sequence are put together in response chains. • Each response produces the signal for the next response. • Chains of responses are organized into larger response patterns.

• A complex response like swimming has three major chains:

Modeling

Modeling

• In addition to classical and operant conditioning, there is a third type of learning called modeling which is observing and imitating imitating.

– An arm stroking chain – A breathing chain – A leg-kicking chain

• It is often necessary learn simple responses before mastering the complex pattern



There are three types of modeling: 1. The simplest case, observing others’ behavior increases the chance that we will do the same thing. g Others clap, p, we clap. p Others look up, we look up. This does not involve learning of new responses, but in using old responses that we are currently not using.

14

10/20/2009

Modeling

Albert Bandura

2. Observational learning. We observe and imitate. Ex.: watch someone perform dance steps and then imitate to learn the steps ourselves 3. Disinhibition – When an observer watches someone engage in a threatening activity without being punished, the observer is more likely to not be inhibited from the same activity. May eliminate a phobia.

Behavior Modification • The systematic application of learning principles to change people’s actions and feelings. • Modeling, Modeling classical classical-conditioning conditioning and operant-conditioning are used to modify behavior.

Computer-Assisted Instruction • CAI is based on operant conditioning, guiding the student through instruction by breaking down the instruction into “frames frames.” As the student masters a concept, he or she is positively reinforced in the form of new information.

Computer-Assisted Instruction • Greek philosopher and teacher Socrates taught his students by a conversational method, similar to what is used today in CAI. CAI • CAI is a refinement of a concept of programmed instruction first introduced by Pressey in 1926 and refined by B.F. Skinner in the 1950s.

Token Economies • Cohen & Filipczak, 1970 – A group of boys labeled “uneducable” were placed in a token economy where they were rewarded with points for good grades on tests. They could “cash in” these points for snacks, lounge privileges or ordered items from a mail-order catalog. Within a few months the IQ of the boys improved an average of 12 ½ points.

15

10/20/2009

Token Economies

Token Economies

• Miller and Schneider, 1970 – Used a token economy to teach preschoolers in a Head Start program to write. The children received tokens they could use for food, movies and other rewards. The children in the token economy improved dramatically whereas those not in the token economy made little progress and showed a less postive attitude toward school.

• In token economies, people are paid to act appropriately. • In overcrowded mental hospitals, the only way some patients can get attention is by “acting crazy.” Most attendants do not have time to bother with people not causing trouble. In effect, people are rewarded for undesirable behavior.

16