Using AI to Understand Dog's Barks and Beyond

TapTechNews June 7th. Researchers are attempting to use AI to interpret whether a dog's bark indicates playfulness or anger. At the same time, researchers are also attempting to use AI to identify a dog's age, gender, and breed.

Researchers from the University of Michigan have collaborated with the National Institute of Astrophysics, Optics, and Electronics (INAOE) in Puebla to conduct this research and found that the AI model originally used for training human speech can serve as a starting point for training models for animal communication.

Rada Mihalcea, head of the University of Michigan's AI laboratory, states that AI has made significant progress in understanding the subtleties of speech, being able to distinguish subtle differences in pitch, tone, and accents, and these research bases can be relied upon to understand dog barks.

One of the main obstacles in developing such an AI model for analyzing animal vocalizations is the lack of publicly available data. While there are many resources and opportunities to record human speech, collecting data from animals is more difficult.

The team tried to collect dog bark information in the same way as collecting human voice data, collecting barks, growls, and whines from 74 dogs of various breeds, ages, and genders in various situations.

The team used these collected sound information in a machine model for analyzing human speech, and the model can well understand the communication between dogs, and the model has an accuracy rate of 70% in various tests.

Rada Mihalcea said, Sounds and patterns from human speech can serve as the basis for analyzing and understanding other sounds, such as animal vocalizations. Other researchers on the team also said that a better understanding of the nuances of the various sounds animals make can improve humans' interpretation and response to their emotional and physical needs.

It is reported that the experimental results were presented at the 2024 International Joint Conference on Computational Linguistics, Language Resources, and Evaluation. TapTechNews attaches the paper link:

https://arxiv.org/pdf/2404.18739