2 Doctoral Researchers in Reinforcement Learning
Core
Developing fundamental theory and practical algorithms for continual reinforcement learning to enable autonomous systems to adapt to changing real-world environments without reset or downtime.
Role type
Doctoral researcher (PhD candidate) in Reinforcement Learning
Builds
Novel RL algorithms integrating trajectory-centric optimization and adaptive, context-dependent policies
Domain
Artificial Intelligence / Robotics / Control Theory
Deliverable
research
Required skills
Reinforcement learning, Stochastic processes, Ergodicity theory, Dynamic programming, Markov decision processes, Python programming
Preferred skills
Experience with robot arms or quadruped robots, Background in change detection algorithms
Technologies
Python, Simulation environments, Hardware robots
Responsibilities
Develop theory for trajectory-centric stochastic optimization under non-ergodic dynamics; Implement practical RL algorithms optimizing long-term performance; Develop change detection algorithms for policy adaptation; Implement efficient policy-adaptation algorithms with safety guarantees; Integrate advances into state-of-the-art RL algorithms; Validate algorithms in simulation and hardware experiments.
