Fabien Couthouis

Fabien Couthouis

Data Scientist - Ubisoft Player Analytics, Montreuil, France

Biography

Interests

  • Machine Learning
  • Deep Learning
  • Natural Language Processing
  • Deep Reinforcement Learning
  • Explainable Artificial Intelligence
  • Human Factors
  • User Experience (UX)

Education

  • MSc in Cognitive Engineering with a major in Artificial Intelligence, 2017 - 2020

    Ecole Nationale Superieure de Cognitique [ENSC] (last year in Ecole Nationale Supérieure d'Electronique, Informatique, Radiocommunications et Mathématiques-Mécanique [ENSEIRB-MATMECA]) - Bordeaux INP

  • Two-years intense Mathematics and Physics curriculum ("CPGE"), 2015 - 2017

    Cycle Préparatoire de Bordeaux (CPBx)

Skills

git

Git

python

Python

c++

C++

unity

Unity

tensorflow

Tensorflow

pytorch

Pytorch

Experience

 
 
 
 
 

Programming Interfaces for a Proprietary Technology (internship)

Thales AVS France

Feb 2020 – Jul 2020 Mérignac, France
I conceived and developed tools to facilitate the use and the integration of a proprietary AI system based on Genetic Fuzzy Trees, which makes use of fuzzy logic and genetic algorithms and allows to learn to an autonomous agent from experiences. The tools I developed simplify the use, integration, diffusion and explainability of this technology. The programming interfaces I developed provide compatibility with common reinforcement learning interfaces. At the end, I participated in a proof of concept in collaboration with other teams: an avoidance module for autonomous drones designed and trained using the tools I developed.
 
 
 
 
 

Automation of Insurance Business Processes (internship)

Covéa

May 2019 – Aug 2019 Niort, France
I first designed and realised proofs of concepts around natural language processing for automatic redirection of client emails: data processing, tuning and testing of both deep learning models and a proprietary expert system. I presented my work as well as an introduction to machine learning to teams not expert in AI. I also participated to a collection of needs for the fraud detection service, by conducting user observations. Finally, I developed a tool to perform user testing for a Voicebot, and I participated in the design of the decision tree behind the Voicebot.
 
 
 
 
 

Assessment of a Cognitive Impairment Screening Solution (internship)

Instituts des Métiers de la Santé, HACS lab, CHU Bordeaux

May 2018 – May 2018 Bordeaux, France
I collected sources of improvement from occupational therapists and students who have used a software designed to detect cognitive impairments. Thus, I conducted and analyzed interviews, user testings and focus groups. Then, I wrote a report on the user experience, summarizing their reviews and including suggestions in order to improve the software.

Recent Publications

Explainability in Deep Reinforcement Learning

A review on explainability in Deep Reinforcement Learning (published in Knowledge-Based Sytems).
A. Heuillet, F. Couthouis, N. Díaz Rodríguez
PDF Source Document

Projects

Some of the projects I realised. More on my Github page!

Virtual Reality Game

Developped using Unity game engine and report on virtual reality with a focus on Human Factors.
PDF Code Slides

GoGo

AlphaZero reproduction for the game of Go with simplified rules.

Smart Mirror

SmartMirror designed for students containing a preconfigured installation of both MagicMirror and vocal assistant Kalliope.

Constraint Binairo

Script to generate binero grids with solution using contraint programming and OR-TOOLS SAT solver.

Classmail

Mail classification Python library optimized for french mails in the field of insurance.

Computer Reversi

Computer othello program implementing an AlphaBeta player with heuristics and a Monte Carlo Tree Search player.

RL Pokemon

Reinforcement Learning environment for Pokemon Showdown battles simulator (agent still in progress).

Accomplish­ments

Practical Deep Learning for Coders

FastAI Sep 2020 – Oct 2020

Hackathon: Make Use of Inter Mutuelles Assistance Data

Inter Mutuelles Assistance Jun 2019

Traitement automatique des langues naturelles

Université de Sherbrooke Feb 2019 – Apr 2019

Deep Learning

Kaggle Jan 2019

Intermediate Machine Learning

Kaggle Dec 2018