[EN]
A research paper authored by our PhD student, Alexandre Lambert, titled ‘Artificial intelligence modelling human mental fatigue: A comprehensive survey’, has recently been published in the Neurocomputing Journal. This work represents a collaborative effort at the intersection of Computer Science and Neuroscience and is part of the interdisciplinary research program (PI-ECE) at the LyRIDS research laboratory.
This achievement marks the first interdisciplinary scientific collaboration that brought together researchers from LyRIDs at ECE and the three research laboratories-, LISV at the University of Paris Saclay, and IRBA (Institut de recherche biomédicale des armées) and LyRIDS.
This collaborative work also involved students from ECE and resulted in another contribution to the JSAN journal [https://www .ece.fr/2023/11/08/mental-fatigue-detection-using-physiological-signals-and-machine-learning /]
The paper subject is described below:
Mental fatigue is a decline in cognitive abilities resulting from prolonged mental effort. It is commonly studied by neuroscientists who understand cognitive aspects and brain chemistry. Yet, defining mental fatigue is still an open question. While neuroscientists have made progress in understanding fatigue, computer scientists, with a limited understanding, have developed models for fatigue detection.
Artificial intelligence offers potential solutions, using fuzzy rules and learning algorithms to predict mental fatigue accurately. However, current models often prioritize acquiring parameters over studying and validating them, leading to reliability issues. Balancing parameter acquisition, validation, and interaction is crucial for accurate mental fatigue models. Our study identifies issues in four areas-experimental design, data processing, parameter choice, and reasoning-and provides practical guidelines for more realistic modeling.
[FR]
A research paper written by our PhD student, Alexandre Lambert, entitled ‘Artificial intelligence modelling human mental fatigue: A comprehensive survey’, was recently published in the journal Neurocomputing. This work represents a collaborative effort at the crossroads of computer science and neuroscience, and is part of the interdisciplinary research program (PI-ECE) at the LyRIDS research laboratory. This achievement marks the first interdisciplinary scientific collaboration between LyRIDS researchers and three research laboratories: LISV (Université Paris Saclay), IRBA (Institut de recherche biomédicale des armées) and LyRIDS.
This collaborative work also involved ECE students and produced another contribution to the JSAN journal. [https://www .ece.fr/2023/11/08/mental-fatigue-detection-using-physiological-signals-and-machine-learning /]
The subject of the article is described below:
Mental fatigue is a decline in cognitive capacity resulting from prolonged mental effort. It is commonly studied by neuroscientists who understand cognitive aspects and brain chemistry. However, defining mental fatigue remains an open question. While neuroscientists have made progress in understanding fatigue, computer scientists, with limited understanding, have developed fatigue detection models.
Artificial intelligence offers potential solutions, using fuzzy rules and learning algorithms to accurately predict mental fatigue. However, current models often prioritize the acquisition of parameters over their study and validation, leading to reliability problems. Balancing parameter acquisition, validation and interaction is crucial for accurate mental fatigue models. Our study identifies problems in four areas: experimental design, data processing, parameter selection and reasoning, and proposes practical guidelines for more realistic modeling.
Updated 2 January 2024