.png)
When we say that a dog has developed odour recognition, we're saying that odour has been classically conditioned; it has gone from being a neutral smell, to a magic smell that predicts the production of all the dog's favourite rewards.
I see a lot of questions about the pros and cons of various different methods of creating odour recognition, so decided to give it a shot to summarize the issue.
Paired OdourThe target odour is paired with the dog’s reward. When the dog finds his reward, he should notice the strong scent of essential oils. Over time, the dog should learn to value the odour in the same way he values food.
Pros:
The dog learns to work independently of handler influence throughout their foundation (pairing) stage. The handler is not involved in the reward sequence until the dog develops value for the odour, which should ideally prevent a loss of independence once the dog learns his human is the keeper of their reward.
Pairing induces the dog to the correct training aid. This means that right from the outset of training they’re learning ‘which of these is not the like the others?’ when there are multiple identical training aids (all sterile bar the ‘hot’ aid). This is an essential point of difference to the other methods, because the dog is always having to conduct a search prior to sourcing odour. As the searching component of detection is arguably the most challenging for dogs, this constant reinforcement of searching is really invaluable.
Cons:
By allowing the dog to self-reward, we have no influence on the behaviour they’re offering at source. Although it’s unlikely that at the pairing stage they’ll offer destructive behaviour (unless the paired odour is housed in an enclosed box), when the handler later becomes responsible for rewarding them, the dog is free to rehearse whatever they please at odour source. In the absence of a marker, they have no clarity as to what behaviour we’re really looking for. There’s a common saying in dog training; Pavlov is always on your shoulder. Well, in this case, so is Skinner. If the dog thinks that scratching and biting is what made the box open, or their human appear with their reward, they’ll keep doing it. When a handler attempts to pre-empt their dog’s destructive behaviour by producing the reward earlier, they tend to create a dependency and cueing issue where the dog notices them hovering ever closer as they near source.
The biggest source of concern when deciding whether to pair or not is the fear that we’re teaching dogs to source something other than the target odour. This is particularly worrying because food is used as a distraction in trial. However, a simple proofing step in the training program will easily mitigate this problem. Plus, even dogs that haven’t had odour paired will need the exact same proofing anyway, as food/toys are very appealing!
Marker Training
When the dog sources odour, we sound a pre-conditioned positive reinforcer; a marker (verbal or clicker, alternatively for deaf dogs, a tactile stimulus is used). After very few repetitions, the dog learns that the target odour predicts the marker, thereby conditioning odour as it’s own reinforcer.
Pros:
Clarity. When the dog sniffs odour, gets marked then rewarded, there’s never any doubt in his mind what produced the magic sound. It takes very few repetitions of this for odour recognition to occur.
More clarity. Again, skinner is on your shoulder. While we’re conditioning odour recognition, we also have the opportunity to create and reinforce the criteria for what will later become a response. If we’re reinforcing the highest concentration of odour as the criteria for the sounding of a marker, that’s exactly where the dog will go every time. When training the focused response, this is the most basic criteria that must be met before adding duration to the freeze behaviour.
Cons:
“Do you really know what you’re marking?” How do we know that we’re not just teaching our dogs to poke something with their nose? How do we know that we’re really marking at the highest concentration of odour? We can’t see odour, after all. There are two points here which warrant further discussion.
Firstly, we can’t see odour. But we can utilize a training aid that produces a reliable odour source. A box, PVC pipe or anything else that has one hole for odour to breathe from creates just that. Doing it any other way would be like trying to teach a dog to go a target in pitch black darkness.
Secondly, there is a concern when using some systems of marker training that we may unintentionally be reinforcing a behaviour, rather than creating odour recognition. In the comic, you can see that the man comes across a red cube. There are no other red cubes to be seen, so he has no choice but to investigate this one, sole cube. This is a common way for some trainers to use a marker in detection. Unlike with pairing, there is nothing that is inducing our man to investigate one particular cube. There is no differentiation taking place. Detection is differentiation. When we teach a dog to poke a lone tin or stick their head in a lone PVC pipe when there’s odour inside, the dog may not be making the connection between the magic smell and the magic sound. This is especially true for all dogs that have free-shaping experience, as they tend to focus more on their behaviour rather than relevant stimuli.
Reward OrientationGood things happen at the odour source. Through repetition, the dog learns that odour predicts those good things. This is a very uncommon approach for the Nosework trainer to use by itself. Reward orientation is normally used in conjunction with pairing and/or marker training. However, for the sake of discussion, it will be discussed as a lone method.
Pros:
Done correctly, it encourages independence. The dog truly believes that the odour source makes his reward fall from the heavens.
Done correctly, it can create an anticipatory response; a momentary pause in movement which – if captured – can be built upon to create a focused response.
Cons:
The dog needs to have a good deal of independence from the outset of training. If they ever happen to check on you and catch you in the act of tossing them their reward, you’ve created a big problem.
The dog needs to be environmentally sound, at least when using a toy/ball reward. All animals have an innate aversion to having things tossed at their head. We’re producing the dog’s reward from behind them, so it needs to travel over their head, plus hit and rebound correctly off source. That could be loud. We could miss and hit them in the head. We could miss and hit a coffee pot (ask me how I know this). Which brings me to the next point…
It takes a tremendous amount of skill and practice. There are alternate systems that utilize a remote-delivery reward system or a scent wall with a trainer hiding behind it that take less skill to use, but these have other drawbacks too (relating more to generalization of searching behaviour and the response).
Which is Best?
Using pairing, marker training and reward orientation to teach odour recognition all have obvious strengths and weaknesses. Personally, I use all three in conjunction with one another:
Inaccessible pairing is used to induce the dog to correct training aid. He is thereby having to differentiate from the beginning. He is never allowed to access the food kept with the target odours. So, he doesn't develop any destructive habits.
A pre-conditioned marker is used to create a positive association with odour, in addition to establishing the primary criteria that will be built upon when the response is the focus. The issue of differentiation is avoided because pairing has allowed us to utilize multiple visually similar, but sterile, training aids.
Reward orientation - not tossing food, but simply delivering it at source - is used to keep the dog near the hide so we can get repeated opportunity to mark/reward per search.
It would be fair to say, however, that the correct use of a marker is by far the best single way of establishing odour recognition. When you think of odour recognition as classically conditioning a target odour, the choice is simple.
#Nosework #NoseworkTraining #OdourRecognition #MarkerTraining #Scentwork #Detectiondogs #detectiondog #mystickfigurecomicsareawesome
Komentarze