Year: 2024
Type: dissertation
Abstract: Artificiel intelligence is a complex concept that requires large amounts of data, sometimes referred to as Big Data. This thesis describes various artificial intelligence architectures, including Bayesian networks, support vector machines, decision trees and neural networks, as well as recent variants such as recurrent neural networks, convolutional neural networks and transformers, to provide a better understanding of the challenges and limitations surrounding their use. Artificial intelligence offers a significant opportunity for creating innovative diagnostic tools in veterinary medicine. As such, this thesis describes various use cases. Historically, artificial intelligence has been developed for the rural sector using descriptive data, particularly through the use of milking robots. Approaches for companion animals, however, are more focused on assisting veterinarians in their clinical decision-making. Artificial intelligence can therefore be used to provide early diagnosis of several diseases and to provide insights when assessing a radiography or a blood analysis. However, studies are often limited to proofs of concept and more research may be needed to achieve the desired results. Acknowledging the inherent challenges associated with the use of artificial intelligence is crucial. Careful management and regulation are required to address primary concerns regarding the way data is used for training and its origin. Effective integration of artificial intelligence into daily veterinary practice while ensuring ethical and high-quality care will also require close collaboration between veterinarians and pet owners. Veterinary artificial intelligence also presents significant opportunities for human medicine and its use should aim for a unified One Health approach.
Source: HAL
Author: Didelot, Hadrien
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7076956563