• Corpus ID: 235765359

Deep Learning for Two-Sided Matching

  title={Deep Learning for Two-Sided Matching},
  author={Sai Srivatsa Ravindranath and Zhe Feng and Shira Li and Jonathan Ma and Scott Duke Kominers and David C. Parkes},
We initiate the use of a multi-layer neural network to model two-sided matching and to explore the design space between strategy-proofness and stability. It is well known that both properties cannot be achieved simultaneously but the efficient frontier in this design space is not understood. We show empirically that it is possible to achieve a good compromise between stability and strategy-proofness—substantially better than that achievable through a convex combination of deferred acceptance… 

Figures from this paper


Deep Learning for Multi-Facility Location Mechanism Design
This work develops a computational procedure for learning almost strategy-proof mechanisms that are as good as or better than benchmarks from the literature, such as the best percentile and dictatorial rules.
Optimal auctions through deep learning
This work overviews recent research results that show how tools from deep learning are shaping up to become a powerful tool for the automated design of near-optimal auctions auctions and recovers to a high degree of accuracy essentially all known analytically derived solutions for multi-item settings.
Automated Mechanism Design via Neural Networks
This paper puts forward a unified neural network based framework that automatically learns to design revenue optimal mechanisms and theoretically proves that the mechanisms found by the framework are indeed optimal.
Certifying Strategyproof Auction Networks
This work proposes ways to explicitly verify strategyproofness under a particular valuation profile using techniques from the neural network verification literature, and trains the network and produces certificates in several settings, including settings for which the optimal strategyproof mechanism is not known.
Automated Mechanism Design without Money via Machine Learning
This work uses statistical machine learning to develop methods for automatically designing mechanisms in domains without money to design strategy-proof mechanisms for social choice problems with single-peaked preferences, and stable mechanisms for two-sided matching problems.
Incentives in Large Random Two-Sided Markets
It is shown that in a probabilistic setting where the preference lists on one side of the market are composed of only a constant number of entries, each drawn from an arbitrary distribution, the number of participants that have more than one stable partner is vanishingly small.
Deep Learning for Revenue-Optimal Auctions with Budgets
This work model a mechanism as a neural network, and use machine learning for the automated design of optimal auctions, and extends the RegretNet framework to handle private budget constraints and Bayesian incentive compatibility.
Strategy-Proofness versus Efficiency in Matching with Indifferences: Redesigning the NYC High School Match
The design of the New York City (NYC) High School match involved tradeoffs among efficiency, stability and strategy-proofness that raise new theoretical questions. We analyze a model with
Stable Matchings, Optimal Assignments, and Linear Programming
It is shown that those fractional matchings in the Stable Marriage Polytope form a lattice with respect to a partial ordering that involves stochastic dominance, and all expected utility functions corresponding to the same ordinal preferences will agree on the relevant comparisons.
The impossibility of strategy-proof, Pareto efficient, and individually rational rules for fractional matching
It is shown that strategy-proofness, ex post Pareto efficiency of the matching, and a weak version of ex ante individual rationality are incompatible when each agent's utility is a linear function of both their fractional assignment and money.