Learning Theory 2.2.pdf

BASIC PRINCIPLES TO REVISE Reinforcers What animals like, what they’ll work for – food, toys, praise, park/walk What do you think are effective reinforcers? Reinforcers are ‘needs’ Secondary Reinforcers – clickers, “good”, picking up leads etc. (represent the reinforcer or indicate that...

BASIC PRINCIPLES TO REVISE Reinforcers What animals like, what they’ll work for – food, toys, praise, park/walk What do you think are effective reinforcers? Reinforcers are ‘needs’ Secondary Reinforcers – clickers, “good”, picking up leads etc. (represent the reinforcer or indicate that it is coming). Lion recall training (Swedish Zoo) – different levels of primary reinforcement (small bits of meat vs. largecarcass feed of something they love), and one secondary reinforcer (3 whistle blows). What about punishers? What they want to avoid? E.g. withholding a reinforcer, withholding your attention (ignoring) How do you know? Test for it, observations, prior knowledge? - Sometimes you need both e.g. recall training – can’t wait forever! (lace monitors, lions) Reinforcer or punishment need to be delivered quickly after the behaviour occurs Why?  BASIC PRINCIPLES - SHAPING Involves reinforcing certain behaviours over others Shaping normally involves two components: a) Reinforcing steps toward the behaviour called successive approximations. Quick use of secondary reinforcer is very useful here! b) Non-reinforcement of earlier behaviour Variability of behaviour is important (meaning it is easier when the animal offers lots of behaviours); Can't reinforce ‘no behavior’ (e.g. to get a dog to stop barking they first need to bark, so you can reinforce stopping). The more behaviour there is the easier it is to select the target behaviour. E.g. a dog giving non-target behaviours helps define what you are reinforcing. The dog need to decide about the following: How do I get more treats? More attention? More play with toys? Get to the park? Get to that patch of grass to smell? Animals don’t ‘know’ what behaviours to give, or ‘know’what we are thinking – they LEARN that they receive a reinforcer, or punishment, after certain behaviours. Their aim is to get more of the reinforcer / less of the punishment (by trial and error to start with). We often use the word ‘know’! ABC’S OF OPERANT CONDITIONING A–B–C Antecedent (what happens immediately before e.g. a cue, stimulus, trigger etc.) Behaviour Consequence What do you think is going to E.g. A = Owner says ‘sit’ happen when the owner says ‘sit’ again? B = Dog sits C = Dog is given a treat for sitting  OPERANT CONDITIONING Links back to the Law of Effect (Thorndike and his cat box) Video from Moodle (7.24 for Operant) If a response is followed by a ‘satisfying’ state of affairs the response is more likely to occur again. = REINFORCEMENT = BEHAVIOUR KEEPS GOING If the response is followed by an ‘annoying’ state of affairs, the response is less likely to occur again = PUNISHMENT = BEHAVIOUR STOPS For example, running a red light might result in you getting to class on time – a good outcome, so you are mor likely to do it again. However, if a cop was behind you then running a red light might lead to a bad outcom (and you would still be late for class). Meaning you would be less likely to do it again next time. OPERANT CONDITIONING Reinforcement and punishment can be: Positive – adding, giving, presenting Negative – subtracting, removal, avoidance Tips Identify the behaviour you want to increase/decrease Identify the best type of consequence  TYPES OF REINFORCEMENT Reinforcement always Now I can GIVE something good = POSITIVE REINFORCEMENT Giving something good e.g. gift Or.. I can take something bad away = NEGATIVE REINFORCEMENT Taking away something bad e.g. discomfort  TYPES OF PUNISHMENT Punishment always Now I can GIVE something bad = POSITIVE PUNISHMENT Giving something bad e.g. verbal punishment Or.. I can take it away = NEGATIVE PUNISHMENT Taking away something good e.g. kids toys/ipad  POSITIVE REINFORCEMENT ADDING something to INCREASE a behaviour Delivering a reinforcer (pleasant) following the target behaviour This makes it more likely that behaviour will occur again and faster. E.g., Reinforcing a dog with a treat for sitting on command  NEGATIVE REINFORCEMENT REMOVING something to INCREASE a behaviour Behaviour is followed by removing something bad This makes it more likely that behaviour will occur again. This is an escape response Courtesy of www.supe E.g., Training a dog to heel with a choker chain. If the dog is slow to heel the choker will tighten The dog will stop pulling and heel to avoid Perform the correct behaviou discomfort. and the bad thing doesn’t happen  POSITIVE PUNISHMENT ADDING something to DECREASE a behaviour Positive Punishment Delivering something bad following the behaviour This makes it unlikely that behaviour Dog chasing tyres, use of a shock collar ONCE resulted will occur again. in cessation of behaviour. Kiwi avoidance for dogs living or working close to habitats E.g., When my dog walks on the road I growl at him. If the growl is aversive to him then the Weigh pros/cons for each behaviour will decrease. situation and Some advocate the use of shock collars for extreme behaviours animal – can anyone think of why?  NEGATIVE PUNISHMENT REMOVING something to DECREASE a behaviour Removal of a (pleasant)consequence following the behaviour This makes it unlikely that Negative Punishment behaviour will occur again. Remove a positive stimulus to decrease unwanted behaviour (ignore by removing yourself from a puppy when he barks) E.g., At the park I throw the ball for my dog. Sometimes before I throw he jumps and barks. I withhold the ball (the object he wants) until he is sitting. I removed a desired object to decrease his jumping and bark