Selected talks (shorter than the papers, and easier to read)

  1. Sequential Decision Making under Parameter Uncertainty(different versions of this talk were given in: Stanford, UT Austin, U of Alberta, MIT, Tel-Aviv U, Technion, Hebrew U 07')./div>

  2. Efficiency Loss in Competitive Market Mechanisms: Asymptotics and Dynamics (Hebrew University, 07).
  3. Calibrated Forecasts: Efficiency vs. Universality (Technion, 06)
  4. An Inequality for Nearly Log-concave Distributions with Applications to Learning (Mathematical Foundations of Learning Theory II, June 06).

  5. Online Learning with Constraints (COLT 06')

  6. Online Learning with Variable Stage Duration (COLT 06')
  7. Decision Making in Dynamic Environments under Parameter Uncertainty (a gentle introduction to MDPs with uncertain parameters, McGill 06)
  8. Reinforcement Learning with Gaussian Processes (slides for a talk that was supposed to be given at the NSF workshop on approximate dynamic programming, 06')
  9. An Inequality for Nearly Log-concave Distributions with Applications to Learning (Hebrew U. 06')
  10. Efficiency Loss in a Resource Allocation Game: Random Users (McGill optimization seminar, Tel-Aviv U., Technion 06')
  11. The cross entropy method for classification (ICML05)
  12. Reinforcement learning with kernels and Gaussian processes (ICML05, workshop on rich representations)
  13. Calibrated Forecasts: Efficiency vs Universality (International Festival on Game Theory 05')
  14. A Contract-Based Model for Directed Network Formation (McGill 05')
  15. Efficiency Loss in a Resource Allocation Game: A Single Link in Elastic Supply (CDC 04')

  16. A Game Theoretic View of Efficiency in Network Resource Allocation (Technion, 04', a more detailed version of the above)

  17. Dynamic Abstraction in Reinforcement Learning via Clustering (ICML 04')

  18. Bias and Variance in Value Function Estimates (ICML 04')

  19. Lower Bounds on the Sample Complexity of Exploration in the Multi-Armed Bandit Problem (COLT 03')

  20. The Cross Entropy Method for Fast Policy Search (ICML 03')

  21. Learning and Adaptation in Competitive Dynamic Environments (Seminar given in Mcgill, UMass, Laval, Duke, and Tel-Aviv University; a more gentle version of the next)

  22. Learning and Adaptation in Competitive Dynamic Environments (Seminar given in Technion, MIT, Brown; proofs are more detailed)

  23. On the Consistency of Boosting Algorithms (MIT)

  24. On the Consistency of Boosting (Haifa Winter Workshop on Computer Science and Statistics, Dec., 2001).

  25. Adaptive Strategies and Regret Minimization in arbitrarily varying Markov Environments (The Fourteenth Annual Conference on Computational Learning Theory, July, 2001)

  26. Geometric Bounds for Linear Weak Learners and Applications to Boosting (Alpine Workshop on Computational Learning, March 2001)