Artificial Intelligence
Research highlights
- New computationally more efficient methods for learning binary Markov random fields
- A formal analysis of finite trace semantics for linear temporal logic in planning
- Foundations of Artificial Intelligence
- Development of a declarative programming language for agent applications
- A new approach to learning partially observable reinforcement learning problems by exploiting sufficient statistics of histories
- Explicit characterisation of convexity of composite binary losses
- New algorithms and theoretical analysis of spatial reasoning which combines topological and directional information for extended spatial objects
- Theoretical analysis of stochastic gradient descent learning algorithms with variable metrics
- New methods for factored planning that allow the efficient (and sometimes optimal) solution of factored planning problems
- A unification of f-divergences, Bregman divergences, statistical information and surrogate regret bounds, cost curves for binary experiments
» Visit the Group's website
» Student research opportunities
