Artificial intelligence methods for the design and simulation of multi-scale epidemiological models in animal health

To speed up and make the development of epidemiological simulations more reliable, we have used artificial intelligence methods to explain the structure of the models and reduce the amount of code to be written. Applied first to intra- and inter-herd models for an animal zoonosis, this generic software approach allows us to cover recurrent needs in modeling communicable diseases in host populations.

Bibliography :

Picault S., Huang Y.-L., Sicard V., Ezanno P. (2017) "Enhancing Sustainability of Complex Epidemiological Models through a Generic Multilevel Agent-based Approach", Proceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI), p. 374-380, AAAI. DOI: 10.24963/ijcai.2017/53

Picault S., Huang Y.-L., Sicard V., Beaudeau F., Ezanno P. (2017) "A Multi-Level Multi-Agent Simulation Framework in Animal Epidemiology", Proceedings of the 15th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS), p. 209-221, Springer LNCS 10349. DOI: 10.1007/978-3-319-59930-4_17

Picault S., Huang Y.-L., Sicard V., Hoch T., Vergu E., Beaudeau F., Ezanno P. (2017, soumis) "A Generic Multi-Level Modelling and Simulation Approach in Computational Epidemiology", Journal of the Royal Society Interface.

Link to the national INRAE blog : http://2025.inra.fr/3perf/Billets/SA-Faits-marquants-2017-contribution-a-OpenScience-3-et-2-et-a-3Perf-2-Des-methodes-d-intelligence-artificielle-pour-la-conception-et-la-simulation-de-modeles-epidemiologiques-multi-echelles-en-sante-animale-TSitre/