From Behavioral Observation to Predictive Ethology: The Role of Artificial Intelligence in Understanding Animal Behavioral

Abstract Study of animal behaviour is an important component of Zoology because animal behaviour has a close relationship with their environment, with other animals, predators, resources and to the changing environmental conditions. Ethological studies are based on direct observation, behaviour sampling, ethograms, video recording of behaviour and hand modified ethograms. So far these methods have led to a lot of knowledge but they are time consuming, observer dependent and do not capture subtle patterns, quick changes in behaviour or complex behaviour. AI's uses in the study of animal behaviour are revolutionised by machine learning, deep learning, computer vision and automated tracking, as well as computational bioacoustics. AI systems can identify people, even guess their physical positioning, track their motion, classify behavioral patterns, analyze voice and body noise, and even identify behavioral patterns that are hard to see with traditional observation. Computational techniques like DeepLabCut, SLEAP, idtracker.ai, JAABA, DeepEthogram, B-SOiD and VAME, have proven the potential of automated behavioural analysis in the laboratory and in the field. However, there are some issues to be faced, including species bias, small sample size, variable behavioural definitions, lack of model transferability, lack of model explainability and lack of field validation. Video, acoustic, physical and ecological data will be integrated with ecological theory in the future and is called “multimodal AI”. AI in behavioural research, critical technologies, applications of AI in animal behaviour, the start of predictive ethology, challenges and future directions of “ethically and biologically responsible” AI development. The programme is designed for artificial intelligence and animal behaviour specialists, computational ethologists, machine learning and deep learning researchers, and anyone interested in predictive ethology.It is aimed at experts in Artificial Intelligence and animal behaviour, computational ethologists, machine learning and deep learning researchers and anybody interested in predictive ethology.