معرفی
Steven Morad is a Lecturer and PhD student in the Department of Computer Science and Technology at the University of Cambridge. His research focuses on reinforcement learning, long-term memory models, and robotics applications, with an emphasis on partially observable environments and multi-agent systems.
He has taught Deep Reinforcement Learning (Lent 2024) and served as a Teaching Assistant for Mobile Robot Systems (2021–2022). His work includes developing frameworks for embodied navigation, memory-augmented algorithms, and cooperative multi-robot systems. Notable contributions include the POPGym benchmark for POMDPs and the NASA-recognized Improving Visual Feature Extraction in Glacial Environments (2019).
Research themes span robotics in extreme environments (e.g., lunar caves, low-gravity terrains) and graph-based methods for partial observability. His publications address challenges in decentralized control, topological priors, and language-conditioned navigation.
- Education: MSc Thesis on The Spinning Projectile Extreme Environment Robot (2019).
- Awards: NASA New Technology (NTR NPO 51401) for glacial robotics work.
- Collaborations: NASA/JPL, IEEE, and aerospace industry partnerships.
