Machine learning, game theory and markov chains
Objectives
-
Understand the details of linear programming problems and their dual versions.
-
Study the applications of linear programming to data science and signal processing problems.
-
Develop the details of game theory necessary to understand the balance of nash in zero-sum games and its relationship to linear programming.
-
Motivate the study of markov chains and their use in machine learning.
Syllabus
1. Linear programming
-
-
Detailed approach and mathematical review
-
Classic applications
-
Applications to data science
-
Comparison with the perceptron
-
Dual programming
-
2. Game theory
-
-
Zero sum games
-
Nash equilibria
-
Mixed strategies
-
Applications of linear programming and von Neumann's theorem
-
3. Invitation to the Markov chains
-
-
Stochastic game theory
-
Formal definitions and probability review
-
Law of the big numbers
-
Time series
-
Sampling methods
-
Ergodicity theorems
-
Relationship with reinforcement learning
-