Sofya (Sofi) Dymchenko

prof_pic.jpg

Hi, salut, привет! I am a Ph.D. student at University Grenoble Alpes and INRIA, I work with Bruno Raffin in Datamove team.

My research is focused on improving the data-efficiency of a deep neural network training on a data which comes from physics simulators and PDEs (usually reffered as surrogates). I am a member of Melissa project which develops a framework for large scale deep surrogate training on HPC systems.

selected publications

  1. Loss-driven sampling within hard-to-learn areas for simulation-based neural network training
    Dymchenko, Sofya,  and Raffin, Bruno
    In MLPS 2023 - Machine Learning and the Physical Sciences Workshop at NeurIPS 2023 - 37th conference on Neural Information Processing Systems Dec 2023
    PDF
  2. MelissaDL x Breed: Towards Data-Efficient On-line Supervised Training of Multi-parametric Surrogates with Active Learning
    Dymchenko, Sofya, Purandare, Abhishek,  and Raffin, Bruno
    In AI4S 2024 - 5th Workshop on artificial intelligence and machine learning for scientific applications Nov 2024
    PDF