research

The foundations of learning and decision-making over time.

We study agents that act, learn from feedback, and carry useful experience from one task to the next. Our aim is to connect rigorous theory with the demands of genuinely adaptive systems.

Current research programmes

ConSequentIAL

Foundations for continual and sequential learning systems.

FoLiReL

Foundations of lifelong reinforcement learning.

Research themes

  • Non-stationarity: detecting and responding to a world that changes over time.
  • Exploration: gathering the right information under partial, delayed, or structured feedback.
  • Risk and control: moving beyond expected return toward reliable decision-making.