Teaching

A reinforcement learning flow diagram illustrating the interaction loop between an Agent and an Environment. A blue arrow labeled 'Action' points from the Agent to the Environment, while orange arrows labeled 'Reward' and state feedback loop back to the Agent.
Reinforcement Learning 

Advanced Computer Science (MPhil)
Lecturer: Rika Antonova 

This course introduces state-of-the-art reinforcement learning (RL) methods, covering theory and practice from foundational concepts to deep RL and applications. Topics include Bellman optimality, DQN and Q-Learning, policy gradient and actor-critic methods, exploration and uncertainty, model-based RL, RLHF for foundation models, and RL for scientific discovery.

A schematic diagram of a deep neural network on a black background. It shows three vertical layers of interconnected nodes (circles) with arrows representing the flow of information and weights between input, hidden, and output layers.
Deep Neural Networks 

Computer Science (Undergraduate)
Lecturers: Ferenc Huszar, Rika Antonova, Nic Lane

This course explores modern neural networks from theory to applications. It covers the basics of optimization (e.g., SGD, Adam), convergence, generalization, automatic differentiation, and mainstream architectures (e.g., CNNs, Transformers). Advanced topics include self-supervised learning, transfer learning, and graph neural networks.

White line-art sketches of three robotic systems on a black background: a quadrupedal robot with a mounted arm, a mobile base with a sensor turret, and a collaborative robotic arm mounted on a wheeled platform.
Advanced Robotics 

Engineering (Tripos Part IIB)
Lecturers: Fumiya Iida, Fulvio Forni, Rika Antonova 

This course builds on introductory robotics and covers advanced topics including underactuated robotics, robot learning, soft robotics, human–robot interaction, and multi-agent systems. Students gain theoretical background and practical experience with collaborative research projects.