Information, Learning, And Incentive Design For Societal Networks
summary
Develops a formal framework marrying game theory, mechanism design, and control to quantify how platform signals change routing and congestion and to design fixes. Proves a complete characterization of Bayesian Wardrop equilibrium, identifies three regimes where information value is positive zero or negative, shows optimal information policies can cut residential overflow without alerting everyone, builds learning tools for non IID data with convergence guarantees, proposes adaptive tolls and a polynomial time market algorithm for carpooling, cites a 83% increase robotaxi mileage result.
AI-written summary · about
chapters (15)
- 0:00Platforms and Users in Societal Systems Al technologies and platforms Information Incentives
- 0:53Societal challenges
- 2:40Disruptive events and security concerns
- 2:59The design questions
- 3:26Approach
- 4:37Outline
- 4:52Part I. Value of information in routing
- 6:20Example: Informed and uninformed
- 7:58Bayesian congestion games
- 8:34Contributions 1. Complete characterization of Bayesian Wardrop equilibrium for
- 9:30Information Design
- 10:17Part II: Multi-Agent Learning with Adaptive Strategies
- 11:39Learning for tolling
- 12:05Part III: Incentive design in networks
- 13:07Summary
related talks
46m(Improving) Muddled Information
44mPlatforms for Efficient and Incentive-Aware Collaboration
01h 12mA Brief Introduction to Algorithms, Game Theory and Risk-Averse Decision Making
01h 00mTutorial: Incentives for Collaborative Learning and Data Sharing, Part Ib
20mAI on the Fast Lane: Smart Traffic Systems to Unclog our Cities
01h 06mLearning in Games II
49mIntroduction and Research Fellows Lightning Talks
30mPolicy Evaluation under Interference