Talk by Johann Benerradi

Johann will talk about his PhD during the lab meeting this week.

Johann Benerradi - Machine learning for mental workload assessment with fNIRS

Mental workload is an interesting concept with stakes on safety, stress and boredom. It has been studied especially in HCI in order to create better user experience. Many ways to assess it have been developed but the most popular rely on questionnaires which are based on subjects perception and are fastidious to answer. This is why we can observe a growing interest in studying the physiology of mental workload to understand its effects and be able to assess it objectively. Among physiology, neuroscience has emerged as a good candidate to study and assess mental workload more precisely, especially with portables devices using functional near-infrared spectroscopy (fNIRS).

My PhD explores ways to assess mental workload from fNIRS data using machine learning, the goal being to build a brain-computer interface capable of giving neurofeedback. I focus more precisely on semi-supervised deep learning approaches to take advantage of labelled but also unlabelled fNIRS data.