Researchers at Mexico’s National Institute of Astrophysics, Optics and Electronics (INAOE) along with the researchers at the University of Michigan, have devised a program using AI to begin the process of understanding dog vocalizations.

They found that AI models used for human speech can be used as a starting point to develop new systems that understand animal vocalizations. The human voice models can encode the very complex patterns of human languages and speech, so the researchers have started to develop similar programs to decode canine barks.

If they are successful, they feel that it will benefit biologists as well as animal behaviorists and other people who work with animals. Of course, it will benefit dogs as well. Such a system could enhance the care of dogs by letting people interpret their emotional and physical needs. This information could prevent potentially dangerous situations.

Sue’s Note: While this is very good, understanding dogs and any other animal depends heavily on their body language. Will this program take that into consideration? To the best of our knowledge, animals do not have a word by word vocal communication system.

Journal Reference:

1. Artem Abzaliev, Humberto Pérez Espinosa, Rada Mihalcea. Towards Dog Bark Decoding: Leveraging Human Speech Processing for Automated Bark Classification. Submited to arXiv, 2024 DOI: 10.48550/arXiv.2404.18739

Cite This Page:

University of Michigan. “Using AI to decode dog vocalizations.” ScienceDaily. ScienceDaily, 4 June 2024. <www.sciencedaily.com/releases/2024/06/240604132204.htm>.