Outline

Active learning
  • Expected utility and decision theory
  • Information theory
  • Linear experimental design
  • Gaussian processes and kernel methods
  • Active learning with GPs
  • Computational learning theory
  • Myopic vs long range active learning
  • Applications to graphics, attention, robotics, web crawling, bioinformatics, ...

    Reinforcement learning
  • Stochastic dynamic programming and control
  • Markov decision processes
  • Bandit problems
  • Stochastic approximation and Q-learning
  • TD-lambda, sarsa, actor critic, value and policy iteration
  • POMDPS
  • Continuous problems

    Undirected graphical models
  • Models
  • Inference algorithms
  • Parameter and structure learning algorithms

    •