DiviyanK I’m a third year PhD student in Université Paris-Saclay, LRI, Inria, working on causality under the supervision of Isabelle Guyon, Michèle Sebag and Philippe Caillou. My background is mostly of Computer Science and Machine Learning.

My main field of research -my PhD subject- is causal discovery out of observational data for social sciences, in the directed graph setting. I’m also interested in :

  1. Generative models
  2. Reinforcement Learning
  3. Computer hardware

I also install/manage clusters under the SLURM batch scheduler, like the 54-GPU cluster at the TAU team. (Also developing python tools to run jobs without trouble. Check out my GitHub repo !)

Publications

Causal Discovery Toolbox

Along with my reasearch work, with Olivier Goudet we are developing a python 3 package for causal discovery out of observational data.

It is still in development, but many of its features are already available. It includes :

  1. Graph utilities and management
  2. Metrics of evaluation for graphs
  3. Many of the pairwise causal discovery algorithms
  4. An extensive list of feature selection and graph recovery methods
  5. Automatic hardware detection (CPU cores, GPUs) and adapting the algorithms to it.

The package is available at : https://github.com/Diviyan-Kalainathan/CausalDiscoveryToolbox