Kevin Denamganaï

Kevin Denamganaï

IGGI PhD Student

University of York

Biography

Good morning (maybe?)! My name is Kevin Denamganaï, I am a PhD Student at the University of York, in the context of the IGGI program. I am advised by Dr James Alfred Walker and Dr Sondess Missaoui.

My main research is in Language Emergence and Grounding to support cooperation, aiming to understand how compositional communication emerges and whether it can be linked to systematicity/(algebraic) generalisation abilities, and be levered towards solving the agent alignment problem.

When I am stuck in this main research, I cope by letting my curiosity loose and so I am usually inquisitive about Unsupervised Representation Learning and Multi-Agent Deep Reinforcement Learning, trying to understand more about the emergence of (situated) cooperation.

Download my resumé.

Interests
  • Unsupervised Representation Learning
  • Language Grounding
  • Language Emergence
  • Deep Reinforcement Learning
  • Basketball
  • Kendo
  • Meditation
  • Astronomy
Education
  • IGGI PhD Student, 2022

    University of York

  • Master Recherche - Intelligence Artificielle et Robotique, 2017

    Université de Cergy-Pontoise

  • Diplôme d'ingénieur - Informatique et Systèmes, 2017

    Ecole National Supérieur de l'Électronique et de ses Applications

  • MEng in Electrical Engineering and Information Science, 2017

    Osaka Prefecture University

Projects

*
Symbolic Behaviour Benchmark (S2B)
Suite of OpenAI Gym-compatible multi-agent reinforcement learning environment to benchmark for behavioral traits pertaining to symbolic behaviours (primarily: being receptive, constructive, malleable, and separable) using referential games.
Symbolic Behaviour Benchmark (S2B)
ReferentialGym
Out-of-the-box Referential Games for Languages Emergence and Grounding in visual modality.
ReferentialGym
GazeboDomainRandomization
This is an implementation of some Domain Randomization tools within the ROS+Gazebo framework, following the work of Tobin et al.’s ‘Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worl’.
GazeboDomainRandomization
RelationalReasoning
State-of-the-art relational reasoning algorithms, using PyTorch.
RelationalReasoning

Recent Publications

Quickly discover relevant content by filtering publications.

Contact