Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore dynamic Bayesian Stackelberg games through this 54-minute Google TechTalk that challenges conventional wisdom about learning in strategic interactions. Examine how leaders can effectively improve their strategies through repeated interactions with fully strategic followers, contrary to existing folk theorems from dynamic pricing literature. Discover novel average-case analysis techniques that demonstrate learning effectiveness without weakening the follower's strategic space, showing that improvement stems from neither commitment ability nor communication substitution. Learn about mixed-integer linear programming algorithms for computing optimal leader policies and explore heuristic approaches for more efficient approximation of dynamic policies. Analyze simulation results comparing algorithm efficiency and runtime against static policy alternatives, gaining insights into market design automation through the intersection of machine learning and algorithmic game theory.
