This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Chapter 03: Fundamentals of Learning
Principles of Learning and Cognition Virginia Commonwealth University drooling dogs cats in boxes rats that press bars In the beginning... Theories built from animal behaviors Believed in "higher mental processing" humans are just advanced animals Trying to develop a general process learning theory for all mammals Two Forms of Learning Eventually developed into two forms of learning... 1.) Pavlovian (a.k.a. classical or respondent) 2.) Instrumental (a.k.a. operant) Examines at "reflexes": involuntary behaviors Examines at "voluntary" behaviors Classical Findings: History Ivan Petrovich Pavlov (18491936) Nobel prize for digestive system work using dogs (1904) Noticed dogs had both learned and innate reflexes From here Pavlovian conditioning took shape
Ivan Pavlov the dog Classical Conditioning Pavlovian style Has a physiological focus Terms to know and understand Unconditioned Stimulus (UCS): stimulus that produces natural/biological response Unconditioned Response (UCR): biological response Conditioned stimulus (CS): stimulus eliciting the biological response through pairing with the UCS Conditioned Response (CR): biological response produced by the CS Reinforcer: another way to understand the UCS since it reinforces the CS to elicit the CR How the CS elicits a response Basic Phenomena of Pavlovian Conditioning Contiguity: if 2 ideas become associated they'll occur in the mind simultaneously Centers in the cerebrum become linked Based on law of contiguity was thought that the UCS and CS should be exactly paired but this didn't work...worked best if CS presented slightly before the UCS and overlaps it If CS begins and ends before UCS called trace conditioning If CS begins well before UCS and overlaps it called delay conditioning Can also pair more than one CS with the UCS to get response...if one CS is more powerful you get overshadowing Conditioning of a distinct positive response: conditioning of excitation Produces minimal conditioning Sometimes inhibition occurs where delayed conditioning is observed... there are several types on inhibition Most important type is extinction After several times presenting the CS w/out the UCS the response will no longer occur The figure shows how the amount of saliva decreases as the CS is not paired with the UCS. Sometimes the response will come back but to a lesser extent. Higher Order Conditioning Definition: when a new CS is built from an old CS...doesn't work well beyond the second level Generalization: Pavlov found that similar stimuli could act to produce the CR Stimulus discrimination: can also teach an organism the difference b/n two potential CS so that only one will elicit a response Ex. tones similar to the original tone presented to the dogs elicited a response Conditioning in Humans Several behaviorists followed Pavlov's ideas to conduct further studies related to humans Tended to not care about the physiology of the brain Only cared about behaviors Pavlov reproved them of this way of doing research later John B. Watson Founder of Behaviorism Worked w/Rosalie Rayner...his lover and future wife Wanted to study basic emotions how conditioning leads to emotions later in life Thought all human behavior was reflexive From this came the idea of conditioned emotional reactions (CERs) Conditioned Emotional Reactions (CERs) The little Albert experiment Paired a loud noise with a rat to make the baby become afraid of rats Then generalized this to many different soft objects and animals Never undid this conditioning but could have used Implosion Counter conditioning modeling Little Albert
UCS: Loud Noise UCR/CR: Fear CS: rat CS: Other soft objects Instrumental Conditioning and Thorndike Life: 18741949 Experiments on Trial andError Learning ... now called operant In this type of learning the subject learns how to operate in their environment to get rewarded Believed in the scientific method so need to invent challenging, controlled environment Came up with puzzle boxes Placed hungry cat or dog in the box The animal had to figure out how to get out of the box Eventually the animal got better at figuring out what to do to get out of the box and get the food Pavlovian Conditioning and the Associative Tradition Stimulus Substitution Theory or Reflex Transfer Account Control of the UCR transfers to the CS when paired with the UCS SR learning occurs b/n the CS and CR... not SS learning b/n the CS and UCS Stimulus Substitution Theory has 5 basic claims that have been modified or abandoned as of now Instrumental Conditioning and the Associative Tradition Thorndike thought that situations and responses are associated...not ideas, memory, etc. Thought learning "consisted only in changing degree of control a stimulus exerts over a response" (p. 74) Three laws of learning proposed 1. Law of Exercise: use of a response strengthens its connection to the stimulus 2. Law of Readiness: If neurons associated with a response are prepared to fire and get to, pleasure will be felt...if inhibited, displeasure occurs 3. Law of Effect: behaviors resulting in pleasure are more likely to occur How many kinds of learning are there? Can't really distinguish Pavlovian and Instrumental based on response produced Can't really distinguish based on procedures either Interaction is shown through autoshaping/automaintenance (p. 77) Overall...could argue there's only one type of learning or several types How Does Learning Proceed? Continuity: learning is a gradual process Discontinuity or insight: learning can occur all at once Thorndike Khler: apes are capable of insight; box andbanana problem E.R. Guthrie (18861959) Looked into how well lab experimentation with animals could relate to real world Believed in singletrial learning (insight) Studied elementary musclemovements in puzzle boxes Voeks continued Guthrie's work No one really argues this topic anymore... sometimes learning is gradual; sometimes it's not For an individual learning occurs through insight As a group learning is gradual (contiguity) What is learned? Latent inhibition: if a stimulus is presented w/out consequence and then used in conditioning, it will take longer for conditioning to occur Distinguishing between knowing how and knowing that Procedural learning: knowing how Declarative learning: knowing that Ex. (p. 81) "most of us know how to ride a bike but most of us do not know the physical principles that we use to do so" The Importance of Context The conditioning context: "background stimuli playing important role in associative learning" Extinction and unlearning are NOT the same thing The UCSCS relationship is not unlearned... the response is just modified Extinction involves performance or expression...not learning (according to Bouton, 1991) Reinstatement of a response to a CS depends on the context Causes of Learning: Pavlovian Conditioning Spaced practice better than distributed practice Short practice intervals with rest in between works better than long periods of practice with no breaks Therefore contiguity is not the sole cause of learning P.S. THIS IS HOW YOU SHOULD BE STUDYING FOR ALL YOUR CLASSES!!!! (50 minutes studying followed by a 10 minute break is recommended) Wow! These people sure look happy while they study!! Pavlov cont. Blocking Unblocking: need to pair the new CS with a new UCS for this to occur (p. 83) More evidence that learning is not just through contiguity Previous and continued pairing of the first CS blocks the formation of a newly paired CS RescorlaWagner model Contingencies affect learning CS must predict the UCS Comparator theories: state that the "organism always learns the CSUCS association even with blocking" Says that with blocking the organism will learn the CS sometimes predicts the UCS RescorlaWagner says no way Instrumental: Learned Helplessness Situation at Time 1: Two dogs are subjected to shocks. One can stop the shocks; the other can't. At time 2: the dog that couldn't stop from getting shocked is put in new environment where he can escape but doesn't when the shock comes...that's learned helplessness Determined that reinforcement is based on contingency b/n the behavior and reward Has implications for depression TX Opposite of this is the illusion of control: believing you are completely in control of your life and now for a short learning break...a moment with Calvin and Hobbes on education What's the Role of Reward? "reward strengthens the SR bond that precedes it" Thorndike Premack Principle: an activity the is engaged in more frequently can be used as a reinforcer for a behavior that occurs less frequently Ex. letting a child go outside and play only after she makes her bed Role of Reward cont. Reinforcer: anything that increases the probability of a behavior's occurrence Schedule of Reinforcement: schedule by which a reinforcer will be received Response deprivation hypothesis (Allison, 1989): can use a less frequent behavior to reinforce a more frequent one Bliss point: the point at which two behaviors are performed at the ideal level p. 87 used example of a rat running and licking to emphasize all these concepts Role of Consciousness Brewer (1974) wanted to know "why" behavior changes Thought that when a behavior changes the subject has "figured out" the experiment and acts accordingly Concluded that there's "no evidence of Pavlovian or instrumental conditioning in adult humans"...or restated..."no conditioning w/out awareness" CRUCR Equivalence: CR as Adaptation to Expected UCR The response is seen as an "adaptive" response as the organism prepares for the UCS that follows the CS Ex. if an bell (CS) warns that a shock (UCS) is coming it seems logical that heart rate (CR) would increase Has been used to explain drug addiction (p. 91) Equipotentiality: Evolutionary Heritage Affects of Associative Learning Assumption of equipotentiality: "any stimulus an animal can perceive has equal potential to become a CS paired with an arbitrary UCS" However...each organism responds differently based on evolution Ex. a fat avoids foods based on taste...not appearance SymbolSystem Hypothesis: Pavlovian Conditioning in a Production System Holyoak, Koh, and Nisbett (1989): how memory systems work...differences from previous model Learning occurs through two processes: "rule generation and changing production strengths" Short term and long term memory events occur...not just procedural "if condition X, then expect Y and take action Z" Elimination of "inference engine"...there is competition for a production to fire during a trial Connectionism: A Model of Pavlovian Conditioning
CS1 Sensory Inputs CS2
Potential Connection Potential Connection UCS Innately Active Connection UCR/CR Output Layer Response Input Layer Connectionism: A Model of Pavlovian Conditioning If the CS or UCS fires then response occurs Configurational Learning: learning that requires multiple and contextual stimuli Negative patterning: most sophisticated form of configurational learning; the reinforcer paired w/ one of the two CS but withheld when both CSs are presented www.markstivers.com/ cartoons/main.html ...
View Full Document
- Spring '05