DOGS could soon be fully understood by their owners — after an AI translator developed a ruff understanding.
A machine designed to transcribe human speech was adapted to decode canine sounds — effectively teaching an old bot new tricks.
And the results left scientists yapping with delight — as it was able to identify different breeds from their barks, and could detect the tone of almost every woof, whine or growl.
Researchers captured more than 200 minutes of barks from 74 dogs in Mexican cities Tepic and Puebla.
The hounds, all family pets aged between five months and seven years, were all recorded in their home environments for authenticity.
Tests found that the bot could successfully differentiate between a variety of stimuli, including angry barking at a dangerous stranger and normal woofs at someone they did not know.
READ MORE PET NEWS
Protective sounds after an assault on its owner were also established, as well as the playful doggy noises during games with a loving family member.
The AI bot, named Wav2Vec2, correctly identified the breed — mostly Chihuahuas, poodles, and Schnauzers — from their bark two-thirds of the time.
The results could pave the way for a greater understanding of man’s best friend, which the study team claims will improve animal welfare.
They also suggested the bot could be trained on other species such as birds and dolphins.
Researcher Rada Mihalcea said: “This is the first time that techniques optimised for human speech have been built upon to help with the decoding of animal communication.
"Our results show the sounds and patterns derived from human speech can serve as a foundation for analysing and understanding the acoustic patterns of other sounds, such as animal vocalisations.”

