Smell and Tell: Lexical and Cognitive Benefits of Long-Term Olfactory Training in Perfumery Students

Anne-Lise SAIVE - Research Scientist - Cognitive Science
Data Science

Date: -
Location: Eurecom

Abstract: This presentation will explore how Natural Language Processing (NLP) can be used to understand human chemosensory abilities, focusing on a project that examines the impact of a year-long olfactory training program at ISIPCA Perfumery School. The study highlights how students converge into using a common specialized vocabulary for describing scents, crucial for their future expertise. NLP and Machine Learning tools were used to assess changes in students' descriptive language and cognitive performance, revealing increased lexical diversity and consistency with expert references. At the end, I will briefly introduce a project analyzing wine descriptors, exploring their relationship to wine quality, types, and countries using NLP. Bio: I am a cognitive neuroscientist with a PhD in Cognitive Neuroscience and five years of postdoctoral experience in Computational Neuroscience. Currently, I am a tenured research scientist at the Lyfe Institute research center in France and a visiting researcher at the Champalimaud Research Foundation in Lisbon, Portugal. My research focuses on odor perception and memory, as well as its interactions with other senses like in food and wine tasting. I use brain imaging (EEG/MEG) and AI-related tools (ML/NLP) to explore odor mental representations in humans. I develop multidisciplinary projects where I collaborate with diverse experts, including perfumers, chefs, and wine specialists. Passionate about naturalistic research, I use immersive experiments including AR and VR, aiming to advance our understanding of human olfaction.