01 July 2016

RESEARCH - Can a Computer Train a Dog?

The world of automation has impacted every industry, and the pet industry is certainly not exempt – from automated pet feeders and exercise equipment…to algorithmic trainers?

According to researchers at North Carolina State University, the better way to ensure consistency in dog training (for example, timed rewards for good behaviour) is to utilize technology to ensure accuracy and improve efficacy of the training process.

(NC State University)
While the idea is not to replace the human trainer with a robot, university researchers Dave Roberts and Alper Bokzkurt are studying the employment of sensors and algorithms to assist the human/dog communication process to aid training and avoid unnecessary human error.

Their project, announced last year, helps handlers monitor guide dogs using a harness and vibrating handle by developing an algorithm that feeds a reward (treat) to the dog in training at the same time, every time; the idea is that this would assist the novice dog trainer, allowing for a smoother training regime with less human error.

For Calgary area-based dog trainer, Rory O’Neill, the science behind algorithms may simply be too far removed from the keys to successful training: trust, common sense and consistency.

“After reviewing online about algorithms, robot dog training, I still feel, it’s simply another bad human idea of how to understand and work with animals,” said O’Neill.

“Animals are not hard to understand, you just need to have compassion and common sense…Bottom line, if anyone wants to resort to teaching their dog with a robot, they instead need to go out and purchase an inexpensive clicker – a training tool used in reward-based training - and spend a few minutes a day encouraging their pet with learning new things.”

O’Neil stresses that the bond between humans and dogs ‘comes very naturally’, but that this bond is not required in the training process – which can instead be based on a fun and fair reward system.

Her personal training methods do not incorporate technology, rather are reliant on consistency, vocal tones, reward systems and above all – not using human psychology to train dogs.

“The mistake is pet dog owners often treat their dogs as children and not meeting the dog’s needs.”

For project lead, Roberts, this research would serve to aid the novice trainer and improve the training process between dog and owner.

“We will never get rid of humans in this process. The human animal bond is really important to all the things we do,” Roberts said. “But just like technology gives humans new ways to interact with each other, it can also give humans and animals new ways to communicate.”

From a perspective of animal welfare, more effective training methods would also increase the adoptability of shelter dogs, according to Roberts, who is hopeful this system would serve the greater good by resulting in better trained dogs whose owners are less likely to surrender their pets to shelters teeming with poorly trained dogs.

By Lindsay Seewalt
Lindsay is an experienced journalist and mother of three whose heart and home is always open to a four-legged friend. With her Corgi, Angie, as household editor-in-chief, Lindsay gives back to the animal planet through the written word on anything and all ado about pets. She is passionate about topics regarding animal welfare and responsible pet ownership, which she aims to instill in both her readers and children to be compassionate animal lovers who are conscious and considerate that furry friends around the globe deserve a voice.

No comments:

Post a Comment