Description

I am a PhD candidate at the University of Lille in the Inria SequeL team, advised by Prof. Olivier Pietquin and Jérémie Mary. My research project explores how to enable computers to acquire language through multimodal interactions. Therefore, I focus my work on goal-oriented visually grounded dialogues. I am both developing new neural architectures to fuse vision and language modalities, and I am extending deep reinforcement learning methods to fit the dialogue setting. Finally, I am also studying dialogues from a multi-player Markov game perspective to develop new training algorithms.

All publications

Workshops

I put a lot of effort in organizing quality workshops to federate communties around research directions. Therefore, I was involved in the organization of the following workshops:

• Reinforcement Learning Summer School (2019) - website
• Visually Grounded Interaction and Language (2018) - website
• European Workshop of Reinforcement Learning (2018) - website
• Visually Grounded Interaction and Language (2017) - website
• International conference of Machine Learning (2015), Local commitee - website

Ph.D Visiting

• DeepMind: March 2018 -> August 2018
• University of Montreal, MILA: June 2017 -> December 2017

Websites

• GuessWhat is a two-player visual dialogue game we developed. We used it as a testbed for Visually Grounding Language Machine Learning algorithms. website
• LightingTorch is a website I set up when I first started to use Lua-Torch. Its initial goal was to provide helpful tips on the Lua Torch Framework. website

Public presentations

• Reinforcement Learning Summer School, Language Learning thought Reinforcement Learning (2019) - website
• ECCV workshop on Visual Learning and Embodied Agents in Simulation Environments (2018) - website
• Machine Learning Meetup Nantes (2017) - The GuessWhat?! storyline - website
• Meet-up Recsys Presentation (2016) - Using Neural Networks to predict user ratings - website - video

Recent blog posts