Automatic safety constraints selection for Safe RL
Work on safe reinforcement learning, focusing on automatic selection of safety constraints, e.g., deactivate constraints for grasping.
Research
Submit a topic →Learning safety constraints for Safe RL
Work on safe reinforcement learning, focusing on learning safety constraints to enforce safety using a model-based approach (the D-ATACOM algorithm).
MushroomRL: an open-source RL library
Contribute to the development of MushroomRL, an open-source Python Reinforcement Learning library, designed for research and education.
CreminiRL: a safe RL library
Port CreminiRL, a Safe Reinforcement Learning library, to Mushroom2.0.
Reinforcement Learning Project
Work on a small Reinforcement Learning project, where you will apply existing algorithms to a simulated environment.
Agentic multiverse analysis
Work at the frontier of reshaping the scientific method for the AI era. You will build efficient ways to statistically explore open-ended research directions with LLM agents and dynamic programming.