{ListeTraductions,#GET{ListeTraductions},#ARRAY{#LANG,#URL_ARTICLE}} {ListeTraductions,#GET{ListeTraductions},#ARRAY{#LANG,#URL_ARTICLE}}
 

Marylou Gabrié has been awarded the L’Oréal-Unesco grant For Women in Science

The purpose of these grants is to highlight the women contribution to science, which is too often kept in silence. Women represent today 28% of the researchers and only 3% of the scientific Nobel Prizes are awarded to them.

« To get over prejudices and inspire vocations », the L’Oréal Foundation, in partnership with Sciences Academy and French National UNESCO Commission, has been committed for 10 years through its L’Oréal-UNESCO Scholarship Program for Women and Science.
These scholarships are awarded each year to 30 doctoral and post-doctoral students selected among more than 900 other candidates for the excellence of their file, the originality of their scientific project and their desire to pass on their passion to the youngest. They will each benefit, in addition to a research grant of € 15,000 for doctoral students and € 20,000 for post-doctoral students, support to enhance their research and give them the visibility they deserve.

JPEG - 85 ko

Marylou Gabrié, PhD student at LPS, is interested in solving the problem of so-called deep learning algorithms.
These algorithms have revolutionized artificial intelligence in the last decade. Inspired by neural networks, they learn to multiparameter programs very complex tasks from a database of examples. However, the mechanisms of this learning remain misunderstood, particularly because of the number of considerable amount of adjusted parameters. Current algorithmic solutions are based on empirical considerations and are resource-intensive : computing power, storage capacity, and so on.

During her PhD, Marylou was studying statistical physics used for the study of systems made up of billions of molecules. "We are exploiting recent advances in this field to study deep learning. This original interdisciplinary approach allows to combine physical approximations and numerical experiments. »
The outcome of this research could in the future bring a better return understanding neural networks of artificial intelligence and thus improve their performance and reliability. Their environmental impact could also be optimized by reducing the computation time, which implies a significant power consumption.

Caroline Rossi-Gendron and Jessica Guerand, Ph.D. students respectively at the Laboratory P.A.S.T.E.U.R of the Department of Chemistry and the Department of Mathematics and Applications (DMA) are also scholarship holders 2018.

Portfolio

  • Marylou Gabrié has been awarded the L'Oréal-Unesco grant For Women in (...)