Personal details |
Name |
Dr David
Leslie |
Job title |
Visiting Fellow
|
Department |
School of Mathematics University of Bristol
|
Personal web page |
http://www.stats.bris.ac.uk/~madsl |
Contact details |
This expert can be contacted via the University of Bristol Public Relations
Office.
To help us deal with your request, please mention the Directory of Experts
when contacting the Public Relations Office.
work+44 (0)117 331 8092
email: public-relations@bristol.ac.uk
|
Qualifications |
M.A.(Cantab.), C.A.S.M., Ph.D.(Bristol) |
Professional details |
Membership of professional bodies |
Council Member of the Royal Statistical Society Member of the Institute of Mathematical Statistics
|
Keywords |
multi-agent learning
learning in games
stochastic approximation
game theory
reinforcement learning
Bayesian modelling
Dirichlet process
sequential Monte Carlo
|
Areas of expertise |
Multi-agent learning
Mathematical methods for coordinating independent agents through learning. In particular, reinforcement learning and fictitious play techniques to ensure convergence to appropriate behaviour when there is minimal information available to the agents.
Bayesian mixture modelling
The Dirichlet process in Bayesian semi-parametric models. Bayesian clustering using mixture models. Mixture models for latent variables. Markov chain Monte Carlo techniques.
Sequential Monte Carlo
Applying sequential Monte Carlo algorithms to real statistical problems. Developing adaptive sequential Monte Carlo algorithms.
|