Clearing Matching Markets Efficiently: Informative signals and Match Recommendations
summary
Models a two-sided market with firms' utility split into observable public scores and unobserved private scores, and agents learn via choice-function queries. It constructs a communication-efficient deferred acceptance protocol that provably recovers a stable matching with high probability using about sqrt(n) communication (up to polylog factors) by adaptive firm thresholds (z_j), selective solicitations, and blocking hopeless applications; also gives a two-round tiered protocol with ~log^2 n signals and preserves DA incentive guarantees.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 0:28Title
- 2:37Research Agenda
- 3:33Model
- 4:20Stable Matching
- 5:44Previous impossibility result
- 7:13Contribution
- 7:55Related work
- 10:52Separable markets
- 21:49Main idea
- 27:33Probability
- 28:38Insights
- 30:20Limitations
- 31:33Horizontal Heterogeneity
- 33:04Tiered Random Markets
- 34:11Unbalanced IID Market
- 38:53How would one implement this
- 40:45How to allow both sides to signal
- 42:37How to prevent hopeless applications
related talks
29mClearing Matching Markets Efficiently: Informative Signals and Match Recommendations
50mDiversity in Matching Markets
56mMarket Design and Computer-Assisted Markets: An Economist's Perspective
36mThe Intractability of Hylland-Zeckhauser and its Aftermath
32mDynamic Matching Models
52mDashboard Mechanisms for Online Marketplaces
50mQuality Selection in Two-Sided Markets: A Price Discrimination Approach
50mStrong Truthfulness in Multi-signal Peer Prediction with Overlapping Tasks