SVL: Goal-Conditioned Reinforcement Learning as Survival Learning
Franki Nguimatsia Tiofack, Fabian Schramm, Théotime Le Hellard, Justin Carpentier — arXiv preprint, 2026
Education
- Ph.D. in Robotics (2023 – 2026) — Inria Paris, ENS
- Expected graduation in summer 2026
- M.Sc. in Robotics, Cognition, Intelligence (2018 – 2021) — TUM
- Graduated with distinction
- B.Sc. in Mechanical Engineering (2014 – 2018) — TUM
Work Experience
- Research Engineer (Jan 2022 – Jan 2023) — Inria Paris
- Research Engineer (Jan 2021 – Jan 2022) — Wandercraft
- Reinforcement learning for self-balancing exoskeleton.
- Output: Patent US20250162141A1.
Research Internships
- Student Researcher (Apr 2020 – Dec 2020) — TUM
- Dynamic Vision and Learning Group
- Supervised by Prof. Laura Leal-Taixé.
- Visiting Researcher (Sep 2019 – Mar 2020) — University of Tokyo
- Nakamura Lab
- Supervised by Prof. Yoshihiko Nakamura.
Publications
Guided Flow Policy: Learning from High-Value Actions in Offline Reinforcement Learning
Franki Nguimatsia Tiofack*, Théotime Le Hellard*, Fabian Schramm*, Nicolas Perrin-Gilbert, Justin Carpentier
* Equal contribution — Fourteenth International Conference on Learning Representations (ICLR), 2026First-order Sobolev Reinforcement Learning
Fabian Schramm, Nicolas Perrin-Gilbert, Justin Carpentier — Differentiable Systems and Scientific Machine Learning, EurIPS 2025, 2025
Reference-Free Sampling-Based Model Predictive Control
Fabian Schramm, Pierre Fabre, Nicolas Perrin-Gilbert, Justin Carpentier — IEEE International Conference on Robotics and Automation (ICRA), 2025
KernelSOS for Global Sampling-Based Optimal Control and Estimation via Semidefinite Programming
Antoine Groudiev, Fabian Schramm, Éloïse Berthier, Justin Carpentier, Frederike Dümbgen — American Control Conference (ACC), 2025
PROXQP: an Efficient and Versatile Quadratic Programming Solver for Real-Time Robotics Applications and Beyond
Antoine Bambade, Fabian Schramm, Sarah El Kazdadi, Stéphane Caron, Adrien Taylor, Justin Carpentier — IEEE Transactions on Robotics, 2025
End-to-End and Highly-Efficient Differentiable Simulation for Robotics
Quentin Le Lidec, Louis Montaut, Yann de Mont-Marin, Fabian Schramm, Justin Carpentier — arXiv, 2024
Leveraging augmented-Lagrangian techniques for differentiating over infeasible quadratic programs in machine learning (Spotlight ICLR)
Antoine Bambade, Fabian Schramm, Adrien Taylor, Justin Carpentier — Twelfth International Conference on Learning Representations (ICLR), 2024
Leveraging Randomized Smoothing for Optimal Control of Nonsmooth Dynamical Systems
Quentin Le Lidec, Fabian Schramm, Louis Montaut, Cordelia Schmid, Ivan Laptev, Justin Carpentier — Nonlinear Analysis: Hybrid Systems, International Federation of Automatic Control (IFAC) journal, 2024
Reactive Stepping for Humanoid Robots using Reinforcement Learning: Application to Standing Push Recovery on the Exoskeleton Atalante
Alexis Duburcq, Fabian Schramm, Guilhem Boéris, Nicolas Bredeche, Yann Chevaleyre — IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022