Platforms for Efficient and Incentive-Aware Collaboration
summary
Shows collaboration as a graph where each node’s payoff needs one unit from its closed neighborhood, and a central platform knows the hidden node permutation while agents only know the graph. Using Bayesian persuasion the platform sends persuasive signals of 0 or 1−ε (randomized with p, ε ≈ 1/√n) to steer actions; a linear-program characterization and a dominating-set + maximal independent-set construction prove a worst-case O(√n) improvement in total workload (tight on double‑star graphs), and a second scheme guarantees per‑instance strict improvement over full disclosure.
AI-written summary · about
chapters (7)
- 0:00Collaborative learning platforms
- 1:25Efficiency and incentives
- 5:45Collaboration networks and stability
- 16:03Bayesian persuasion for incentives
- 27:20Improving price of stability
- 38:48Summary and open problems
- 41:38Q&A
related talks
01h 00mTutorial: Incentives for Collaborative Learning and Data Sharing, Part Ib
45mTutorial: Incentives for Collaborative Learning and Data Sharing, Part I
14mInformation, Learning, And Incentive Design For Societal Networks
01h 07mIndustry Presentations
48mPersonalized Collaborative Learning with Affinity-Based Variance Reduction
01h 27mTutorial: Incentives for Collaborative Learning and Data Sharing, Part II
30mLearning Multi-Agent Collaborations With Decomposition
01h 04mNonparametric Bayesian Methods: Models, Algorithms, and Applications IV