Petpuls Lab Inc. has harnessed AI to achieve perhaps the first instance of unidirectional interspecies communication made possible through machine learning. Through a Fitbit-sized dog collar, the algorithm collects ‘voice’ data of your dog, telling if your best friend is happy, anxious, relaxed, angry or sad. Petpuls’ proprietary ML algorithm utilizes its database of more than 10,000 bark samples collected from 50 diverse breeds in determining the emotional state via the AI collar.
SwissCognitive Guest Blogger: Abhinav Raj, a political correspondent for Immigration Advice Services.
What’s in a bark? For starters, data. AI-powered dog collar from South Korean start-up can analyse a wide range of emotions—from your dog’s barks. A South Korean wearable tech start-up is showing the world that translating your dog’s barks is paws-ible with AI.
Petpuls Lab Inc. has harnessed AI to achieve perhaps the first instance of unidirectional interspecies communication made possible through machine learning.
Through a Fitbit-sized dog collar, the algorithm collects ‘voice’ data of your dog, telling if your best friend is happy, anxious, relaxed, angry or sad. Petpuls’ proprietary ML algorithm utilizes its database of more than 10,000 bark samples collected from 50 diverse breeds in determining the emotional state of the dog wearing the collar. The collar pairs with Petpuls smartphone application available on iOS and Android operating systems. The app offers insight into the dog’s emotional status and physical activity.
How does it work?
Machine learning and deep learning methods have been employed in the past to audio analysis to classify, describe or interpret data from audio samples. These audio samples are represented graphically across an amplitude-time graph. The audio samples then, by comparative study, provide valuable insight into the emitter—which in this case is a dog.
The pet collar analyses the dog’s bark and through matching and cross-referencing voice data through its database, it computes the temperament of the dog, indicating whether its happy, anxious, resting, at ease, or agitated.
The beauty of machine-learning based models is that the more they are utilized, the better they get at predictive modelling. With every application, the AI dog collar gets better at determining the mood of the dog.
The collar boasts of an 80 per cent accuracy rate in the determination of the emotional state of the dog, according to a research study by Seoul National University.
“This device gives a dog a voice so that humans can understand,” said Andrew Gil, director of global marketing at Petpuls Lab, in a statement to Reuters. “More people began to adopt dogs, but unfortunately some of them abandoned their dogs due to miscommunication. Petpuls can have an important role in the pandemic … it helps owners understand how dogs feel and increases their bonding.”
AI is progressing in all avenues of life. The AI collar brings humanity a step closer to achieving interspecies communication and understanding the species that dwell alongside us better. In decoding the languages of the species that share the planet with us, we move forward in decoding the language of nature.
The development is pioneering in bridging the communication between pets and their human counterparts, which given time, may become more commonplace with more diversified applications across animal species through ML algorithms—hinting of a future where it’s possible to have a chat with our pets to know what’s up. Humanity may soon have its own Doctor Dolittle—made possible through machine learning.
About the Author:
Abhinav Raj is a political correspondent for the Immigration Advice Service, a UK-based organization of immigration solicitors that provides Indefinite Leave to Remain (ILR) services, Visa assistance for prospective migrants and pro-bono legal counsel.