Leveraging artificial intelligence in bioacoustics for animal health monitoring and early diagnosis in veterinary medicine.

Abstract

The study of animal communications, termed zoosemiotics, includes the sub-field of bioacoustics, the study of the production, transmission, and reception of animal sounds. It is becoming increasingly apparent that inter- and intra-species communication is sophisticated with sound playing a major role in this signaling. Artificial intelligence-led research can be employed to understand and combine recorded multi-level data (sound, vision, odors) to classify animal health and identify interventions, also determining critical time-points for intervention. This can include subgroup discovery and trajectory analysis as essential elements in developing animal specific identification of failure to thrive or ill health. It is important that animals, carers, and veterinarians receive as early a diagnosis as possible to predict trajectory and plan care needs and interventions. However, the use of quantitative data for evidence-led interventions based on sound have not yet been developed. Here we look at advances in bioacoustics and provide a framework to determine where early diagnosis and animal health improvements can be made via understanding of behavior and oral sound production.