Mathieu Chalvidal Email & Phone Number

Ville De Paris, Île-De-France, France

Research in deep reinforcement learning, meta-learning and optimization. Thesis title: Learning higher-order functions for computation and control in artificial and biological neural networksAbstract: After decades of research, the question of how biological neural networks synthesize lived experience to serve higher-level cognitive processes (concept acquisition, systematic decision making, evaluative thinking, creativity...) still resists a complete scientific understanding. Concurrently, advances in large scale computation have determined the development of their artificial counterparts and enabled to break new ground in machine intelligence. These simplified neural models present however evident behavioral limitations compared to biological neural networks and provide unsatisfactory to little insight into the computational principles that evolution condensed in biological neurons. One particular limitation is the pour adaptability of current artificial neural systems to external parameters, determining their current failure as a viable functional programming paradigm. In an attempt to identify key mechanisms able to potentially power more abstractive artificial systems, I explore the problem of learning higher-order functions (i.e functions of functions) with neural networks. In particular, inspired by theories from functional analysis and neuroscience, we will consider the specific higher-order problem of learning to learn (or meta-learning). We show that considering neural networks as controllable functionals able to design their own learning strategies with minimal solution constraints as a function of data, neural activity or choice history, can help build new adaptive neural programs displaying better robustness and generalization properties at lower training sample complexity, with broad application to physical system identification, robotic motor control or generative modeling.Multiple NeurIPS papers (2021, 2022, 2023 spotlight) ICLR (2020)

Région De Sydney, Australie

Design, implementation and automation of ready-to-use analytical solutions to support business growth over different thematics from core platform optimisation to marketing insights retrieval. Principal missions included A/B testing, marketing/core data exploration and visualisation, machine learning building (Natural language processing, image recognition...) Tools: Tableau, R, Python (Scikit-Learn, keras, Tensorflow), SQL, Teradata, Hadoop

Paris

Agorize is a French Start-Up specialized in open innovation consulting supporting big companies (Air Liquide, Aviva, Total, Groupe BPCE...) My mission consisted in managing project of Agorize client portfolio: Plannification, Accountancy, Creativity workshop animation, Acceleration and incubation programm survey...

Région De Paris, France

Rédaction d'une étude préliminaire à la refonte de la Marketplace digital du groupe Dassault Sytèmes. - Veille stratégique et benchmark des solutions concurrentes- Réalisation d'enquêtes de satisfactions (focus group, entretiens clients...)- Modélisation du site- Recommendations ergonomiques et graphiques

Msc In Artificial Intelligence, Mathematics And Computer Science

Activities and Societies:Course curriculum included: - Deep Learning (in Theory and Practice) - Reinforcment Learning - Foundations of Big.

Diplôme D'Ingénieur Des Arts Et Manufactures, Ingénierie Généraliste

Major in Applied Mathematics with a specialisation in Data Science. Course curriculum included: - Stochastic process (Ito process, Wiener.

Master Of Science In Management, Technologies De L'​'Information, Statistiques, Comptabilité, Gmat : 670 - Toeic (Listening And Reading) : 950

Activities and Societies: Essec Rugby Team - Pao Bang (agence de graphisme professionelle étudiante) - Mardis de l'ESSEC (tribune.