Titre : Human assessment of artificial intelligence results: Experts feedback on the prediction of their patients? trajectories outcome

Sujet proposé dans : M2 MOSIG, Projet --- M2 MSIAM, Projet

Responsable(s) :

Mots-clés :
Durée du projet : 5 mois
Nombre maximal d'étudiants : 1
Places disponibles : 1
Interrogation effectuée le : 07 mars 2021, ŕ 05 heures 03



Artificial Intelligence advances now allow effective prediction of various outcomes of patients medical trajectories. Designing such prediction system requires a considerable interdisciplinary collaborative effort, that includes understanding of the data, of the prediction outcomes, and of the medical professionals perception of the predictions themselves. 

The purpose of this internship is to design a comprehensive framework allowing to include in the prediction loop the expert perception. The framework will leverage on existing paradigms in human validation of artificial intelligence results to allow meaningful and generalizable experiments. Such experiments will be conducted to validate and feed further advances in the prediction systems. 



References : 

Rodrigues-Jr, J. F., Spadon, G., Brandoli, B., & Amer-Yahia, S. (2019). Patient trajectory prediction in the Mimic-III dataset, challenges and pitfalls. arXiv preprint arXiv:1909.04605. https://arxiv.org/pdf/1909.04605.pdf

Rodrigues-Jr, J. F., Pepin, J. L., Goeuriot, L. and Amer-Yahia, S. (2020). An extensive investigation on Machine Learning techniques for apnea screening. In CIKM Applied Research.