Learning And Conditioning PDF

LEARNING AND CONDITIONING COURSE OF EXPERIMENTAL PSYCHOLOGY Università Cattolica del Sacro Cuore Milan CLAUDIA REPETTO TOC — Definition and general concepts — Classical Conditioning (Pavlov) — Instrumental Conditioning...

LEARNING AND CONDITIONING COURSE OF EXPERIMENTAL PSYCHOLOGY Università Cattolica del Sacro Cuore Milan CLAUDIA REPETTO TOC — Definition and general concepts — Classical Conditioning (Pavlov) — Instrumental Conditioning (Skinner) — Shaping — Latent learning (Tolman) — Insight (Kholer) — Observational Learning  Learning is a relatively permanent change in behaviour, that is brought about by the experience. DEFINITION Pay attention….. A change in performance does not necessarily imply learning: - improvements may be due to the maturation of the biological and cognitive systems (nature) (i.e. motor coordination) Pay attention….. A change in performance does not necessarily imply learning: -decreases in performance may be due to fatigue (i.e. sport) CLASSICAL CONDITIONING Pavlov (1849-1936) discovered the classical conditioning mechanism Learning= involuntary association between stimulus and response  Experimental setting This device measured the quantity of salivation produced by the dog after the presentation of different stimuli Pavlov was a physiologist and he wasn’t interested in studying psychology…The discovery of the conditioning was a serendipity. It occured when he observed a curious phenomenon: the dog started producing saliva and stomach acids before the food was provided to it, as it saw the experimenter approaching Classic Conditioning: a type of learning in which a neutral stimulus triggers a response after it has been paired with a stimulus that naturally brings about this response EXPERIMENTAL STEPS At first the dog receives an unconditioned stimulus - UCS (food), which yields an unconditioned response - UCR (salivation), thanks to the natural link between UCS and UCR Unconditioned Stimulus ® Unconditioned Response (FOOD) (SALIVATION) At this point a neutral stimulus– NS (light) must be associated to the presentation of the UCS. After repeated associations between NS and UCS, the NS is able by itself to elicit the response, therefore it becomes a conditioned stimulus – CS NS UCS UCR (Light) (Food) (Salivation) CS CR (Light) (Salivation) The response brought about by a CS is a conditioned response - CR  The phase just described is called the acquisition phase The unconditioned stimulus is regularly preceded by the neutral one, until the conditioning occurs The neutral stimulus becomes a conditioned stimulus  The next stage is called extinction If light is always presented alone, i.e. not followed by food, the salivation response progressively decreases The conditioned stimulus loses its ability to generate the conditioned response After the extinction, and a short rest period (about 24 h) if the dog is presented again with the CS, it will again show a CR Spontaneous recovery No reinforced trials are needed. It demonstrates that the association between CS and CR is still available in memory A further phase called re-acquisition can be carried out The combination between neutral stimulus and unconditional stimulus is re-established, to the point of creating conditioning again. The time needed to establish the new conditioning is much shorter than that taken in the first phase of acquisition The learning curve is steeper Classical conditioning phase chart  Events related to classical conditioning GENERALIZATION DISCRIMINATION same response to similar different responses to stimuli different stimuli Ex: conditioned response Ex: conditioned response to both to white and red light the presentation of light but not to that of a sound — Once the CS is “acquired” it could act as UCS to establish another association with a neutral stimulus Second-order conditioning INSTRUMENTAL CONDITIONING Skinner (1904-1990) introduced the instrumental conditioning model Learning= the consequences of a response modify the likelihood that the response will be produced again in the future  “All we need to know in order to describe and explain behavior is this: actions followed by good outcomes are likely to recur, and actions followed by bad outcomes are less likely to recur.” (Skinner, 1953) Instrumental conditioning has its roots in the Law of effect (Thorndike, 1898) An action yielding positive (pleasant) outcomes is more likely repeated in the future On the contrary a behaviour generating negative outcomes (unpleasant) is less likely repeated again Skinner’s box Accidentally the pigeon will hit the button with its beak and will receive a dose of food from the distributor. After several random behaviours, an association between the button press and the food delivery will be established The pigeon will hit the button on purpose when it wants to receive food (operant behaviour – voluntary)  THE CONCEPT OF REINFORCEMENT REINFORCEMENT = environmental change that follows an operant behaviour It is the process whereby the delivery of a stimulus increases the probability of a behavior. PRIMARY REINFORCER: stimulus that satisfies basic drives, such as the biological ones; it is not related to the previous individual’s experience SECONDARY (CONDITIONED) REINFORCER: stimulus acting as reinforcer thanks to its consistent association with a primary reinforcer  TYPES OF REINFORCEMENTS POSITIVE: Delivery of a NEGATIVE: Removal of pleasant or appetitive an unpleasant or stimulus following a aversive stimulus behavioral response (e.g. following a food) behavioral response (e.g. headache) PUNISHMENT: Delivery of an unpleasant or aversive stimulus following a behavioral response (e.g. electric shock)  What type (s) of reinforcement are featured in this video? [video Giulia] TEST  REINFORCEMENT SCHEDULES Based on the timing with which the reinforcement is administered, we can obtain two different programs: continuous intermittent Reinforcement follows Reinforcement is each valid operant administered behaviour occasionally  INTERMITTENT REINFORCEMENT Is is defined according to the parameter that regulates the intermittence Interval Ratio (time) (number of beh.) FIXED FIXED VARIABLE VARIABLE  Ratio Interval (numbers of (time) beh.) FIXED FIXED (every x (after a fixed sec./min) number of beh.) VARIABLE VARIABLE (after a variable (after a variable amount of number of beh.) sec./min) Piece work Call center Exams Surprise test  WHICH IS THE MOST EFFECTIVE? At the beginning, when learning is to be established, the continuous program is more effective. When learning is achieved, it keeps up well thanks to the intermittent program.  Discrimination as stimulus control training à a discriminative stimulus signals the possibility that a behavior is followed by reinforcement Generalization à behavior that has been reinforced in certain situations is generalized to other situations (prejudice?) Discrimination e generalization In instrumental conditioning Which is the most effective? Reinforcement Punishment  SOME EXAMPLES… Escape learning: moving away from a dangerous or unpleasant situation is the result of negative reinforcement (the cessation of the unpleasant event). Avoidance: learning to avoid punishment by responding with getaway to a stimulus that precedes it. Learned helplessness (Seligman e Meier, 1975): following exposure to aversive stimuli without the possibility of escape, it is much more difficult to learn avoidance behavior  How is it possible to condition an action that cannot occur randomly (since it is not part of the species-specific repertoire of the individual)? SHAPING  Shaping consists in reinforcing behaviors that are progressively more and more similar to the target. Ex: to get the dolphin to jump into the circle, reinforcement is given first when it approaches the circle, then when it goes around it, then when it jumps near the circle, and finally when it jumps into the circle. LATENT LEARNING Tolman (1930) rejects the vision of automatic learning, mediated by the stimulus-response relationship. He believes that a series of mental operations occur between the presentation of the stimulus and the emission of the response in the subject MENTAL REPRESENTATION The rats are placed in the maze and begin to explore the path in search of food. 3 groups: 1. Immediate reinforcement 2. No reinforcement 3. Deferred reincforcement EXPERIMENT This type of learning occurs even in the absence of reinforcement. The rats learn a cognitive map of the maze and elaborate a mental representation of the path. LEARNING BY INSIGHT Kohler (1925) studied the behavior of chimpanzees who had to reach a banana placed outside the cage. They had bamboo poles available, but none long enough to attract the banana. How will the chimps do it? After some time, the chimpanzee grabs two rods and fits them in order to obtain a sufficiently long one. He had a INSIGHT ABILITY TO RESTRUCTURE THE ELEMENTS IN THE ENVIRONMENT IN A DIFFERENT AND CREATIVE WAY IN ORDER TO SOLVE A PROBLEM THAT IS NOT DIFFERENTLY SOLVABLE Insight is a non-associative learning, and does not take place according to the trial and error method For an insight to be made, there must be a coexistence of all the elements of the problem in the environment. Stages: - failure - pause - observation of potential tools - successful attempt, thanks to the sudden restructuring of the perceptual field OBSERVATIONAL LEARNING A. Bandura (1977) studies a type of learning that makes use of the observation of another individual called a model. Learning takes place by imitation THE BOBO-DOLL STUDY — One group of children was shown adult models behaving aggressively towards a Bobo doll — Another group of children was exposed to adult models behaving non aggressively — During free play, children of the first group displayed more aggressive behavior towards the Bobo doll The stages of observational learning Pay attention to the critical features of the model Remember the behavior Reproduce the behavior Be motivated to learn and repeat the behavior in the future Observational learning: genetic background MIRROR NEURONS Observational learning: when? ü when the trial and error approach is not advantageous (surgeon, pilot etc...) ü if the model was rewarded for the action she performed ü if the model has not been rewarded I learn the behavior but I am less inclined to repeat it