Tag Archives: positive reinforcement

Will computers replace dog trainers?

North Carolina State University researchers have developed and used a customized suite of technologies that allows a computer to train a dog autonomously (without human involvement), with the computer effectively responding to the dog based on the dog’s body language.

“Our approach can be used to train dogs efficiently and effectively,” says David Roberts, an assistant professor of computer science at NC State and co-author of a paper on the work. “We use sensors in custom dog harnesses to monitor a dog’s posture, and the computer reinforces the correct behavior quickly and with near-perfect consistency.”

Dog training with computers

“Because the technology integrates fundamental principles of animal learning into a computational system, we are confident it can be applied to a wide range of canine behaviors,” says Alper Bozkurt, an assistant professor of electrical and computer engineering and co-author of the paper. “For example, it could be used to more quickly train service dogs. Ultimately, we think the technology will be used in conjunction with human-directed training.”

The dog harness fits comfortably onto the dog and is equipped with a variety of technologies that can monitor the dog’s posture and body language. Each harness also incorporates a computer the size of a deck of cards that transmits the sensor data wirelessly.

For the current study, the researchers wrote an algorithm that triggered a beeping sound and the release of dog treats from a nearby dispenser whenever the dog’s harness sensors detected that the dog went from standing to sitting.

The researchers had to ensure that the reinforcement was given shortly after the desired posture was exhibited, and also ensure that rewards were only given for the correct posture. This required a trade-off. If the algorithm ran long enough to ensure the correct posture with 100 percent certainty, the reinforcement was given too late to be effective for training purposes. But if the reinforcement was given immediately, there was a high rate of rewarding the wrong posture.

To address this, the researchers worked with 16 volunteers and their dogs to optimize the algorithm, finding the best possible combination of speed and accuracy. The researchers then compared the algorithm’s timing and accuracy to that of an expert human trainer.

The algorithm was highly accurate, rewarding the appropriate behavior 96 percent of the time. But the human trainer was better – with a 100 percent accuracy rate.

However, while the average response time was about the same for both algorithm and trainer, there was a lot of variation in the time of response from the trainer. The algorithm was incredibly consistent.

“That variation matters, because consistency is fundamentally important for all animal training,” Roberts says.

“This study was a proof of concept, and demonstrates that this approach works,” Bozkurt says. “Next steps include teaching dogs to perform specific behaviors on cue, and integrating computer-assisted training and human-directed training for use in various service dog applications.”

“In the long term, we’re interested in using this approach to animal-computer interaction to allow dogs to ‘use’ computers,” Roberts says. “For example, allowing an explosive detection dog to safely and clearly mark when it detects components of a bomb, or allowing diabetic alert dogs to use their physical posture and behaviors to call for help.”

Source:  North Caroline State University media release

Electronic training collars are a welfare risk

Animal behaviour specialists at the University of Lincoln (UK) have published a study that supports the use of positive reward-based training methods over the use of electronic shock collars.

Shock collar

The immediate effects of training pet dogs with an electronic collar cause behavioural signs of distress, particularly when used at high settings.

The study involved 63 pet dogs referred for poor recall and related problems, including livestock worrying, which are the main reasons for collar use in the UK. The dogs were split into three groups – one using e-collars and two as control groups.

The trainers in the study were industry approved and fully familiar with the guidelines for use of e-collars which are published by collar manufacturers.

Trainers used lower settings with a pre-warning function and behavioural responses were less marked than during a preliminary study (the results of which have since been discounted because the trainers did not follow collar protocols). Despite this, dogs trained with e-collars showed behavioural changes that were consistent with a negative response. These included showing more signs of tension, more yawning and less time engaged in environmental interaction than the control dogs.

Following training most owners reported improvements in their dog’s problem behaviour. Owners of dogs trained using e-collars were, however, less confident of applying the training approach demonstrated.

These findings indicate that there is no consistent benefit to be gained from e-collar training, but greater welfare concerns compared with positive reward-based training.

Lead author Jonathan Cooper, Professor of Animal Behaviour and Welfare at the University of Lincoln’s School of Life Sciences, said: “e-collar training did not result in a substantially superior response to training in comparison to similarly experienced trainers who do not use e-collars to improve recall and control chasing behaviour.

Accordingly, it seems that the routine use of e-collars even in accordance with best practice, as suggested by collar manufacturers, presents a risk to the well-being of pet dogs. The scale of this risk would be expected to be increased when practice falls outside of this ideal.”

The peer-reviewed journal article for this research is:

Jonathan J Cooper, Nina Cracknell, Jessica Hardiman, Hannah Wright, Daniel Mills ‘The welfare consequences and efficacy of training pet dogs with remote electronic training collars in comparison to reward based training’ PLOSone http://dx.plos.org/10.1371/journal.pone.0102722

Source:  University of Lincoln media statement

How dogs were trained for functional MRI research

I’m an absolute advocate for positive reinforcement training.  Here’s how Professor Gregory Berns and his research team trained dogs to remain still in a noisy MRI scanner.

You can also read my other blogs about functional MRI research and dogs:

 

Dogs’ brains respond to human voices

Yet more research on how dogs’ brains work.  This time from a research team at Eötvös Loránd University in Hungary and published in the journal Current Biology.

Using functional MRI, the team could see where blood flowed in the brains of a group of 11 dogs.  The dogs had been specially trained using positive reinforcement techniques to lie still in the MRI scanner for six minutes.

A dog lies still in the fMRI scanner, wearing earphones to pipe in sounds as part of the study. (Photo by Eniko Kubinyi)

A dog lies still in the fMRI scanner, wearing earphones to pipe in sounds as part of the study. (Photo by Eniko Kubinyi)

which tracks blood flow to various areas of the brain, a sign of increased activity—to peer inside the minds of dogs. One of a handful of labs groups worldwide that’s using the technology in this way, they’ve used positive reinforcement training to get a study group of 11 dogs to voluntarily enter the fMRI scanner and stay perfectly still for minutes at a tRead more: http://www.smithsonianmag.com/science-nature/your-dog-can-tell-from-your-voice-if-youre-happy-or-sad-180949807/#DXcpTX0jfeQGFWVY.99
Give the gift of Smithsonian magazine for only $12! http://bit.ly/1cGUiGv
Follow us: @SmithsonianMag on Twitter

The team played each dog a series of over 200 sounds across several MRI sessions.  The sounds included human voices, dog vocalizations, and meaningless noises.

When the results were compared, it showed that the dogs’ brains appear to have a dedicated area that displays more activity in response to voices (whether human speech or dogs barking) than other meaningless noises (such as glass breaking).

More importantly, that part of the brain shows more activity upon hearing an emotionally positive sound, as compared to a negative one.  This means that our dogs are able to distinguish a tone of voice that is positive from one that is negative. (Something many of us probably already knew)

The voice areas of the dogs’ brains is similar to that found in humans, suggesting that our species evolved from a common ancestor almost 100 million years ago, enabling a high degree of communication and social structure.

“We know that dogs don’t have language, per se, but we see now that dogs have very similar mechanisms to process social information as humans,” Attila Andics, lead researcher on the study says. “It makes us wonder what aspects of so-called ‘language skills’ are not so human-specific after all, but are also there in other species. That’s something we plan to look at.”

Source:  Smithsonian Magazine

Here are my earlier blogs about functional MRI studies on dogs:

they show that the dogs’ brains appear to have a dedicated area that displays more activity in response to voices (whether human speech or dogs barking) than other meaningless noises (such as glass breaking), and that part of this area shows more activity upon hearing an emotionally positive sound, as compared to a negative one.Read more: http://www.smithsonianmag.com/science-nature/your-dog-can-tell-from-your-voice-if-youre-happy-or-sad-180949807/#DXcpTX0jfeQGFWVY.99
Give the gift of Smithsonian magazine for only $12! http://bit.ly/1cGUiGv
Follow us: @SmithsonianMag on Twitter

Who’s a good girl..and other things we’d like to hear

Have you ever considered that the things you say to your dog are a reflection of what you would like to hear?  Food for thought…

Dentistry without anesthesia?

I had a phone call yesterday from an older gentleman who wanted to know if I cleaned dogs’ teeth without anesthetic.  (I do not)  I told him that I do sell dog toothbrushes and toothpaste and he wasn’t aware that these things existed!  And then he said that he would never be able to brush his dog’s teeth because he wouldn’t cooperate.

This made me wonder why he thought a dental procedure without sedating his dog was going to be effective.  But I try to be positive when engaging with new clientele; it seems that he was after a procedure that would cost him less money – not necessarily what was appropriate (or comfortable) for his dog.

I explained that I thought he would be better served by having a proper dental cleaning which is a veterinary procedure and then focusing on prevention.  This would include things like teeth-brushing by getting his dog accustomed to the taste of the paste, then gradually introducing the brush with praise and treats (positive reinforcement) throughout.

He thanked me for my advice.

I’ve done a little homework about this practice, because I have concerns whether an animal could truly be treated thoroughly without sedation.  I’ve just been to the dentist myself for a cleaning this week and it isn’t always a comfortable procedure!  Imagine a dog being restrained for it…

The American Veterinary Dental College refers to Anesthesia-Free Dentistry as Non-Professional Dental Scaling (NPDS) and cautions owners against the procedure for the following reasons:

1. Dental tartar is firmly adhered to the surface of the teeth. Scaling to remove tartar is accomplished using ultrasonic and sonic power scalers, plus hand instruments that must have a sharp working edge to be used effectively. Even slight head movement by the patient could result in injury to the oral tissues of the patient, and the operator may be bitten when the patient reacts.
2. Professional dental scaling includes scaling the surfaces of the teeth both above and below the gingival margin (gum line), followed by dental polishing. The most critical part of a dental scaling procedure is scaling the tooth surfaces that are within the gingival pocket (the subgingival space between the gum and the root), where periodontal disease is active. Because the patient cooperates, dental scaling of human teeth performed by a professional trained in the procedures can be completed successfully without anesthesia. However, access to the subgingival area of every tooth is impossible in an unanesthetized canine or feline patient. Removal of dental tartar on the visible surfaces of the teeth has little effect on a pet’s health, and provides a false sense of accomplishment. The effect is purely cosmetic.

3. Inhalation anesthesia using a cuffed endotracheal tube provides three important advantages… the cooperation of the patient with a procedure it does not understand, elimination of pain resulting from examination and treatment of affected dental tissues during the procedure, and protection of the airway and lungs from accidental aspiration.

4. A complete oral examination, which is an important part of a professional dental scaling procedure, is not possible in an unanesthetized patient. The surfaces of the teeth facing the tongue cannot be examined, and areas of disease and discomfort are likely to be missed.

In my blog on Managing Dental Health, I explain the things I do for Daisy to keep her teeth in good condition.  I recommend professional veterinary care to ensure your dog’s oral health, followed by a preventative regime that minimizes the need for future cleanings and anesthesia.

Outmoded notion of the alpha wolf

It’s been almost a year and a half since I wrote about the alpha roll myth.  Yet, there are still dog trainers who are using methods that are based on outdated thinking about animal behaviour and training.

Here’s a great video by L David Mech, who wrote “The Wolf: Ecology and Behavior of an Endangered Species” in 1968.  The book, published in 1970 and re-published in paperback in 1981, is often cited as the reason why ‘dominance’ and ‘leadership’ models for dog training are acceptable.

L David Mech now admits he was wrong and has publicly announced on his website that he has pleaded with the publisher to stop publishing his book.

“Alpha” implies competing with others and becoming top dog by winning a contest or battle. However, most wolves who lead packs achieved their position simply by mating and producing pups, which then became their pack. In other words they are merely breeders, or parents, and that’s all we call them today, the “breeding male,” “breeding female,” or “male parent,” “female parent,” or the “adult male” or “adult female.”

In the rare packs that include more than one breeding animal, the “dominant breeder” can be called that, and any breeding daughter can be called a “subordinate breeder.”

I’m a supporter of positive reinforcement training.  Please be on the lookout for trainers who still use outdated information and possibly damaging training techniques.