Tag Archives: learning

06/7/13

The Range of Imitation in Dogs

ResearchBlogging.orgThis is a brief review of the extensive work by German researcher Friederike Range, looking at imitation or modelling behaviour in dogs. Previously, I posted anecdotal evidence on modelling in dogs (which many people shared their experiences on). This post is more sciencey!

Border terrier and young woman running in large paddock with bushes behind them.

Dogs can imitate the behaviour of both dogs and people. Imitation success depends on a range of factors, including:

  • The task at hand, including its complexity, has a role in imitation.
  • If a human is modelling the behaviour, if they are talking, this can either help or hinder the dog’s modelling, depending on the task. Eye contact can also help (or hinder) a dog in a task. It can help, as it may illustrate to the dogs bits that it should pay particular attention to. However, it make hinder as it may distract the dog from the task at hand.
  • Training plays a role. Dogs that are ‘better trained’ are better at making deduction on behaviour from witnessing a model.
  • The dog’s individual personality.

Interestingly, dogs don’t ‘blindly model’ behaviour of other dogs. They will try to be more efficient, they will learn from the other model’s mistakes, and make adjustments based on particular circumstances. For example, if a dog witnesses the model dog carrying a ball and pushing a lever with its foot, the dog will imitate by pushing the lever with its mouth. The dog seems to realise the model dog used its foot because its mouth was occupied. However, if the model dog pushed the lever with its foot without anything in his mouth, the dog will imitate the foot-push. (Almost, the dog imitates superstitious behaviours.)

Fascinatingly, several dogs have been trained to imitate behaviour.  A dog called Joy was trained with the cue ‘Do it!’.  The experimenter would do one of eight behaviours, say ‘do it’, and Joy would then do the behaviour just demonstrated. After several weeks, they then asked Joy to ‘do’ a behaviour that the experimenter had never demonstrated behaviour. Joy did it. Joy understood the concept of ‘do what I do’. You can see more videos of dogs ‘doing it’.

The conclusions are: Yes, of course dogs model the behaviour of people and dogs. Indeed, they can be trained to do so. There is still a lot of research going on about all the facets of imitation, and it’s all truly fascinating. Definitely a space to watch.

 

 

Links of Interest

Do dogs imitate? by Patricia McConnell.

Dogs show human-like learning ability.

If you’re aggressive, your dog probably will be to.

Dogs automatically imitate people.

 

References

Huber, L, Range, F, Viranyi, Z & Voelkl, B 2008, ‘The evolution of imitation: Old wine in new bottles?’ (PDF).

Range, F., Heucke, S., Gruber, C., Konz, A., Huber, L., & Virányi, Z. (2009). The effect of ostensive cues on dogs’ performance in a manipulative social learning task Applied Animal Behaviour Science, 120 (3-4), 170-178 DOI: 10.1016/j.applanim.2009.05.012

Range, F., & Viranyi, Z. (2009). Different aspects of social learning in dogs Journal of Veterinary Behavior: Clinical Applications and Research, 4 (6) DOI: 10.1016/j.jveb.2009.05.007

Range, F., Viranyi, Z., & Huber, L. (2007). Selective Imitation in Domestic Dogs Current Biology, 17 (10), 868-872 DOI: 10.1016/j.cub.2007.04.026

Szucsich, A, Range, F, Miklosi, A, Huber, L 2008, ‘Imitative ability of dogs‘.

Tiefenthaler, M, Range, F, & Huber, L 2008, ‘Personality in dogs and its influence on social learning behaviour‘.

Virányi, Z., & Range, F. (2009). How does ostensive communication influence social learning in dogs? Journal of Veterinary Behavior: Clinical Applications and Research, 4 (2) DOI: 10.1016/j.jveb.2008.10.023

Viranyi, Z, Range, F, & Huber L 2008, ‘The influence of ostensive demonstration on selective imitation in dogs‘.

12/21/11

McGreevy: General Dog Training Thoughts

This post is part of the McGreevy seminar series. Click here for the index.

 

McGreevy described animal training as “a bit of an art and a bit of a science”. ‘Training’ animals means changing the frequency to which animals show certain behaviours. Learning theory is a universal language that clarifies the nature of training, explaining what will work and will not work, and its general principles apply regardless of the species being trained.

Training often seeks to establish connections between two or more events, and does so by using operant conditioning (i.e. rewards and punishments) and classical conditioning, and often these two work together.  ‘Conditioning’ is any relatively permanent response that occurs as a result of exercise (that is, any responses formed by maturation or debility are not from conditioning).

Trainers often have exquisite timing, and have the ability to self reflect on their progress.

 

 

Photos © Ruthless Photos

“Life coaches”

McGreevy prefers to use the term ‘life coach’ to describe the relationship between a dog and a person.  Life coaches have opportunities for the dog to have success, but also rules.  (The concept of ‘alpha’ asks for people to adopt an unrealistic, pseudo dog role that is not very useful for dog training.)  How dogs and people interact is relevant to the dog’s success.  The handler of a dog needs to be relevant to the dog – a boring or passive life coach is irrelevant for the dog, and the dog will not work.  Dogs will form a bond with their owners, and a trust, but this trust is not generalisable to all situations or to different people.

‘Trust’, itself, is an interesting concept.  It is difficult to measure, and is built on consistency.  During training, trust is built be trainers being caregivers and companions rather than ‘leaders’ or ‘dominant’.

Generally in dog training, we seek dogs that will respond to cues (e.g. the word ‘sit’) with appropriate behaviours.  It is an ongoing process that requires maintenance in many contexts and environments.

 

Dog social order

Dogs with one another have a social order, but it’s not so much a hierarchy. Dog social order is built on difference, not dominance.  This ‘difference’ is a different desire for different resources, meaning some dogs are more inclined to seek some resources than others.  The ideas of social order shouldn’t be ‘thrown out’ with dominance theory.  In short, dogs have evolved to compete with one another.  Excellent coaches tap into the resources that dogs compete over, and use them in training (as rewards).

 

This concludes our section on training dogs, but we will continue to investigate more McGreevy topics in posts to come.

 

In the meantime, I wonder:

What do you think are the ‘art’ and the ‘science’ bits of dog training?

How does your self-reflection as a trainer go?

How would you measure trust with your dogs?

 

This post is part of the McGreevy seminar series. Click here for the index.

12/9/11

McGreevy on Classical Conditioning

This post is part of the McGreevy seminar series. Click here for the index.

 

This is a short summary on Classical Conditioning, as mentioned at Paul McGreevy’s recent seminar.  For a more inclusive approach, please see this post from the Reactive Champion blog or Roger Abrantes’ post on “Unveiling the myth of reinforcers and punishers“.

Classical conditioning is also called Pavlovian conditioning.  In classical conditioning, an association is made between two things.  That is, when I hear the letterbox clang, I associate this with getting mail.  When I walk past a bakery, I associate the smell with the taste of bread.  When I see dark grey clouds, I think about the forthcoming rain.  Sounds, colours, smells, pretty much anything, can be associated with other things.

Puppies are more likely to experience novel events and form strong associations. (Picture courtesy of Yorke Peninsula Puppy Rescue.)

Simply, classical conditioning makes the world predictable with “X goes with Y”, where ‘X’ could be anything and ‘Y’ could be anything else.

The more closely X is followed by Y, the more likely the association will be made.

When an animal makes an association (i.e. “X goes with Y”), and it forms an involuntary response, it is has said to be classically conditioned.  Often, trainers feed dog-reactive-dogs in the presence of other dogs in order to have them associate ‘dogs with food’ and, involuntarily, feel more comfortable in the presence of other dogs.

When this association is novel, then the association is more rapidly acquired because there are no ‘undoings’ to be done.  For dogs, puppies that are attacked by a particular breed may associate that breed (appearance, smell, colouration, body language) with being attacked, and have a fearful response as a result of this classical conditioning.  A single bad experience, if novel enough, can elicit a fearful response for life (this is also true of operant conditioning).

Sometimes, dogs may make associations based on slight cues in the environment.  This can lead to them responding in ways that can confuse human counterparts.  Indeed, during any training and any interactions, classical conditioning may occur.  Some are simple: The dog may associate the car with fun.  Some are more complex: The dog associates your brother with thunderstorms, because that was the context on first meeting.

McGreevy provided some insights into other animals and the classical conditioning that is typical in their species.  He described how stallions used at stud often associate particular bridle gear with sex, and so become aroused and otherwise agitated when that bridle gear is brought out.  (My stud dog associates ‘new little wirey dog’ in the house with sex, which is not always great when it’s actually a male foster dog!)  McGreevy also mentioned how dairy cows associate the sound of the vacuum with releasing milk, and sometimes release milk prematurely upon hearing that sound.

Further reading: Ian Dunbar on Classical Conditioning 

 

This post is part of the McGreevy seminar series. Click here for the index.

12/1/11

McGreevy and Non-Associative Learning

This post is part of the McGreevy seminar series. Click here for the index.

 

Paul McGreevy spoke a little about non-associative forms of learning.  I found these particularly interesting, as I have not actively considered these forms of learning before, let alone attended a seminar about them!  Though I had heard of habituation and sensitisation before, McGreevy clarified and solidified my prior understanding.

Mostly, the seminar considered classical and operant conditioining.  In both of these forms of learning, an association is made between one thing and another.  For example, if a dog hears the fridge open, they associate this with food, and start salivating.  Similarly, a dog may associate the action of sitting with a toy reward.

Dogs can learn to habituate to water.

However, training can also be non-associative.  Non associative learning uses a single stimuli (i.e. ‘thing’, trigger).  The two forms of non-associative learning, as McGreevy put it, are habituation and sensitisation.

 Habituation

Habituation is the simplest form of learning.  When an animal is repeatedly exposed to a stimuli, and their response decreases on each exposure, then they are habituating to the stimuli.

For example, a dog may startle at the sound of a door slamming shut in the wind initially.  However, as this occurs once or twice a day, the dog begins to startle less and, eventually, barely responds to the door slams. This dog has been habituated to the door slamming noise.

Habituation can occur at different rates.  It depends on the stimulus, the frequency of presentation, and the regularity of exposure.

The stimulus itself can be benign or more alarming. For example, a rock on the ground is less alarming than a bag blowing in the wind. In this way, the nature of the stimulus (i.e. the stimulus itself) can affect the likelihood of habituation taking place, with ‘boring’ stimuli more likely to be habituated to.

The frequency of presentation of the stimulus can also the rate of habituation.  Clearly, if an animal was exposed only once to an aversive stimuli, they are unlikely to reduce their response in subsequent exposures.  For example, some dogs are alarmed by the smell of predators, like foxes.  A dog exposed to fox smell may at first respond in a hyper-aroused or anxious way.  However, over time, the dog may begin to habituate to the smell of a fox if it was frequent (for example, if the dog was frequently walked along a route with fox smell).

Finally, the regularity of exposure is important.  If an animal is exposed to a stimulus only once a year or once every few months, it is unlikely to habituate.  For example, if the exposure was daily, then they are more likely to habituate to the phenomena.  McGreevy made reference to puppy schools, and how they normally happen on an irregular basis (i.e. once a week) and this is not enough for puppies to become habituated to people and other dogs.

Sensitisation

Sensitisation, on the other hand, is the opposite of habiutitation.  During sensitisation, the animal responds more to a stimulus after repeated exposure.  In this case, the stimulus itself has to be intrinsically aversive.  For example, dogs who are scared of thunderstorms can often get more fearful over time.

Our next posts in the McGreevy series will look at Classical Conditioning.

 

Further reading: Dogs deciding whether something is important or not?

 

This post is part of the McGreevy seminar series. Click here for the index.

01/5/11

Mini-Jackpotting

I’ve always been somewhat sceptical on the concept of jackpotting.  I don’t know why it has never sat well with me – it just seems a bit much to comprehend that dogs can have an understanding of a degree of success.

That being said, my experience does indicate some benefits in jackpotting.  I guess the best description of what I do is ‘mini-jackpotting’.This is what I use when free shaping behaviours, and I reward ‘more successful’ attempts with more food.

Over the last couple of days, I have been training scent identification and indication.  The process was very slow, until I started mini-jackpotting. In this example, my scent was a teabag and I wanted my dog to scratch/dig at the teabag.

Over the session, I was rewarded different interactions in different ways.  My dog would be rewarded with one piece of kibble if the looked at or moved towards the teabag.  I rewarded touching the object with a paw with numerous bits (about 5 pieces). An actual scratch or dig with about 10 pieces.

My dog was very slow at first, but mini-jackpotting seemed to very much speed up the learning process.  There are several reasons that this may be the case…

  1. I read once that dogs understand the time of a reward more than the quantity of the reward… i.e. Dogs find it more rewarding to be given 5 treats in a row, one after the other, rather than being given a handful of 5 treats.  So, dogs find a long reward more rewarding. (Unfortunately I don’t recall the source of this suggestion.)  As it takes more time to eat numerous treats, perhaps the dog understands this as more rewarding.
  2. Another approach on the time front is that when the dog is eating numerous treats, they are actually having time to think. Perhaps when I reward many-treats at once, the dog has more of an opportunity to think through and the improvements I see towards my target behaviour are actually from this thinking time, rather than the reward itself.
  3. The dog might actually understand that if they do x they get more treats than if they do y!

This is the most thought I’ve ever given to ‘mini jackpotting’, and I haven’t been very logical in its implementation.  If this system occurred by accident or subconscious desire to jackpot, I am unsure.  However, I have found it to be quite successful and I would be interested to see if anyone has had similar success.

Further reading: Schedules of Reinforcement