Skip to main content

About me 🔎

I graduated from École des Mines de Saint-Etienne and Imperial College London in 2017. At the Imperial College London, I realized my master thesis on the detection of sources propagating rumours in social networks under the supervision of Prof. Pier Luigi Dragotti .

I worked at Air France Operational Research from 2018 to 2019, doing predictive maintenance on the A380 and A320 fleets.

In december 2023 I defended my PhD, supervised by Hichem Sahbi at Sorbonne University, in the MLIA and then SYEL team of the LIP6. This PhD is realized in partnership with the LIP6 and Netatmo, a company designing smart and connected products for the house. My current research focuses on neural network compression, pruning and sparsity in general.

Publication and Academia 📑 #

  • PhD Manuscript – Sorbonne Université, 12/2023
    Deep Neural Network Compression for Visual Recognition.
    PDF SLIDES (🇫🇷)

  • Extracting Effective Subnetworks with Gumbel-Softmax - ICIP 2022, Oral
    Robin Dupont, Mohammed Amine Alaoui, Hichem Sahbi, Alice Lebois
    ARXIV SLIDES POSTER VIDEO

  • Weight Reparametrization for Budget-Aware Network Pruning - ICIP 2021
    Robin Dupont, Hichem Sahbi, Guillaume Michel
    ARXIV SLIDES POSTER VIDEO

  • Master Thesis - Imperial College London, 2017
    I studied rumours propagation in social networks and social graph by modelling rumour spreading and creating a maximum likelihood detection algorithm for rumour sources. This master thesis was made under the supervision of Prof. Pier Luigi Dragotti.
    PDF POSTER

  • Social Network Analysis Project - University of Edinburgh, 2015
    I built algorithms for automatic extraction and analysis of social networks taken from literary masterpieces.
    PDF

Miscellaneous 🧪 #

I like DIY, and here are a few things I am proud of:

  • I built my own mechanical keyboard running the open-source firmware qmk , my own electric guitar and magic mirror , among other things
  • I retro-engineered the local API of my connected speaker and created a python library for anybody to use, available on github and PyPi
  • I bypassed my ISP box to use my own router and setup several VLANs to isolate IoT devices from trusted devices at home
  • I run several useful services at home such as Nextcloud , Home Assistant , AdGuard or Jellyfin , all in docker containers
  • I created a bot called Turbodose , which sends notifications on your smartphone when there is an available appointment nearby to get vaccinated (Covid19)