Enlightened Equitation
May 23, 2012, 03:26:54 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Articles Login Register Chat Shop Join EE Events 2012 Free DVD  
Fibre Feeds
Pages: [1]   Go Down
  Print  
Author Topic: Terminology related to clicker/positive reinforcement  (Read 1458 times)
0 Members and 1 Guest are viewing this topic.
TashaKat
Guest
« on: November 02, 2008, 10:00:35 PM »

The more I learn about clicker the more fascinating it becomes  nod  What strikes me, though, is that the correct terminology is rarely mentioned except for by people like Becky, Kassidy etc.  I like to understand how things work rather than just following a script so it's good to know more about the science behind it (or maybe I'm just weird) nod

So ... how about a glossary of terminology for those of us who need to know more and want to know more Smiley

I'm sure that most of us will know some of them already but I'm sure that we've also got some gaps  Embarrassed


Please feel free to add and answer Cheesy


« Last Edit: November 03, 2008, 03:26:27 PM by TashaKat » Logged
OliverG
EE Society Member
Hero Member
*****
Offline Offline

Posts: 1646


EETT Level I New Zealand


« Reply #1 on: November 02, 2008, 11:30:16 PM »

I have no idea, but will read with interest!
Logged

*** Dont argue with an Idiot, you will have to stoop to their level, and they will beat you with experience EVERY time!***
kassidy
Guest
« Reply #2 on: November 03, 2008, 02:40:26 PM »

I'll do the basics if you want -

Aversive (stimulus) - A stimulus is a change in the horse's internal environment (e.g. blood sugar or internal pain) or external environment (e.g. new visual information) which is detectable by the horse's 5 senses. An aversive stimulus is any stimulus which the horse would consider to be unpleasent, uncomfortable or undesirable e.g. pressure, pain, hunger etc.

Bridge or bridging stimulus - A bridging stimulus is a stimulus, such as a sound or visual cue, which the horse has learnt means he is preforming the correct behaviour and to keep going. Clicker can be used in the manner - see Ben Hart.

Terminal stimulus - A terminal stimulus is a stimulus which the horse has learnt means he has performed the behaviour correctly and to stop. This is the way most people use clicker such as Alexander Kurland and Becky Holden.

Bridge and target training - Pioneered by Kayce Cover the system uses a signal for bridging and different signal for the terminal stimulus giving the animal the most possible information about whether they are performing the behaviour correctly. http://synalia.com/ A very interesting training technique. Two different clicks could possibly be used for this, Kayce uses xxxxx and X as the bridge and terminal stimulus respectively. 

Chaining - Chaining is the linking of different behaviours to form one continuace chain. Chain is traditional done backwards with the last behaviour being link to the 2nd to last, and then the 3rd to last behaviour is added etc etc. This means that the animal always goes on to perform behaviours it knows are rewarded. Chaining different behaviours can also be done forwards but it relies purely on repetition.

Classical conditioning - Classical conditioning conditioning is a learnt association between two stimuli or events. These can be pleasent or unpleasent associations, for example, human in feed room = food or  saddle = pain (with a badly fitting saddle).

Conditioned reinforcement - Do you mean operant conditioning? Operant conditioning is the association between a behaviour and a consequence. Depending on whether the consequence reinforces or punishes the behaviour, the behaviour will be strengthen or weakened respectively. For example, standing still at mounting block = treat, standing still behaviour reinforced and will become stronger. However if standing still at mounting block = being dug in the side and slammed down on, the standing behaviour is punished and thus weakened.

Counter conditioning - Counter conditioning is when a unwanted behaviour or reaction is extinguished by strengthening (e.g. through reward) a behaviour or reaction that is incompatible to the unwanted one. For example, rewarding clamness to replace fear or anxiety, associating loud noises or anything scary with rewards such as food, or preventing aggression by rewarding friendliness.

Luring - Luring is the use of food to encourage the horse to perform a desired behaviour so that it can be rewarded. Once the behaviour is learnt and cued, the lure is removed. For example, when teaching the horse to come to call at first it is often useful to use food to encourage the horse over.

Primary reinforcer  - Is any reinforcer required for the horse's continued survival, e.g. food and water.

Secondary reinforcer - Is a reinforcer that predicts the arrival of a primary reinforcer, i.e. the click in clicker, the click has no innate value but predicts the arrival of food. Similar to tokens in an arcade predict rewards of prizes or money allows us to purchase those things needed for survival, even though small bits of paper and metal have no interinsic value their association with primary reinforcers gives them value. Secondary reinforcers could be said to be conditioned reinforcers. (Secondary reinforcer can also be any reinforcer the horse finds desirable and pleasent but is not required for there continued survival, e.g. scratches.) 


Hope this is some help. As I say only the very basics.  Smiley
« Last Edit: November 03, 2008, 02:44:07 PM by kassidy » Logged
TashaKat
Guest
« Reply #3 on: November 03, 2008, 03:25:10 PM »

Brilliant Cheesy  Just the job, thank you Cheesy


Logged
amberleystar
Guest
« Reply #4 on: January 14, 2009, 12:14:42 PM »

I am in the process of writing about the various jargon used in clicker training and behaviour work for my website, so here are some more to add to the list - in draft of course  Cheesy

Successive approximation
The process of breaking down the required behaviour into small steps and reinforcing each step towards the finished behaviour. Successive approximation is the scientific term for shaping.

Free shaping
Reinforcing/rewarding naturally occurring behaviours, i.e. not using cues, lures or giving any other type of information signal. The horse has to work out what you want.

Microshaping
As used by Alexandra Kurland - the same principles as free shaping, however you are looking for micro movements from the horse – muscle twitches/tiny movements towards the behaviour you want. Rewarding the smallest try. This can be quite intense work for the horse and you intersperse this micoshaping with targeting – this easier behaviour acts as extra reinforcement to the microshaping behaviour.

Modelling
The use of equipment such as side-reins etc to put the horse in the desired shape. Modelling is really negative reinforcement – using pressure - horse releases – remove pressure and reward with click and treat.

Habituation
Scientifically, habituation can be described as the process whereby an animal becomes accustomed to non-threatening environmental stimuli and learns to ignore them.

Learned irrelevance
Similar to habituation. It is when the animal has learned to ignore your cue, usually due to a lack of understanding, or where a cue doesn’t have a consistent or sufficiently strong significance attached to it. Common with dogs, especially during recall when the come cue has attained learned irrelevance.

Poisoned cues
A poisoned cue is one that sometimes has a negative consequence instead of a positive one.  A cue becomes poisoned when you cue a behaviour and use a correction such as negative reinforcement or punishment for the incorrect behaviour and positive reinforcement for the correct behaviour.

For a behaviour trained entirely with positive reinforcement, if one now clicks for correct behaviour following a discriminator ( a cue, command, or signal) but also gives an aversive correction (negative reinforcement, punishment) for incorrect behaviour following that same stimulus, the stimulus immediately loses its value as a positive reinforcer. It is, at best, ambiguous in terms of reinforcement. It is not a click. It no longer automatically triggers the positive emotions associated with conditioned positive reinforcers. It can no longer be predictably used inside a chain to reinforce previous behaviour.

Logged
Becky holden
EE Teachers
Hero Member
*****
Offline Offline

Posts: 660



« Reply #5 on: January 14, 2009, 09:08:08 PM »

Quote
What strikes me, though, is that the correct terminology is rarely mentioned except for by people like Becky, Kassidy etc.  I like to understand how things work rather than just following a script so it's good to know more about the science behind it (or maybe I'm just weird)

I dont use "terminology" and have been criticised for it in the passed! cc_confused I think people should understand the correct scientific terms, as i do, but i know it can frighten people off trying to teach with positive methods if they become consumed with science to soon.
I follow the principles of teaching using classical methods backed up with the clicker. So im sorry but cant add to the excellent posts from kassidy and amberleystar.

For anybody wanting to have a go with clicker training and you think that correct terminology sounds confusing please dont be put off!! It really isn't as difficult as it sounds!

Tashakat....... I dont think you are weird, there is nothing wrong with wanting to know more and i think its a good idea that this thread has been pinned so people can refer back to it.

Becky Smiley
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  




Graphics by Mandeigh

Powered by MySQL Powered by PHP Powered by SMF 1.1.16 | SMF © 2011, Simple Machines
SMFAds for Free Forums
SMF customization services by 2by2host.com
Valid XHTML 1.0! Valid CSS!