Ryo Shirakawa | Searcharxiv

arXiv2025

Allocating Common-Value Goods

We study a simple problem of allocating common-value goods. The designer seeks to allocate the goods to as many unit-demand agents as possible without monetary transfers, while agents, who possess partial private information about the goods, are willing to receive them only when the goods are of high value. Mechanisms screen each agent's private information using the information of other agents, and in doing so shape what agents learn from other agents about the value of the goods. The optimal mechanism can be summarized by two parameters: one adjusts the allocation probability, while the other governs the amount of learning induced by allocation. Although the designer prefers to allocate the goods, the optimal mechanism excludes some agents and, as a result, may withhold allocation even when all agents would be willing to receive them. The optimal mechanism has the same structure even when payments are available, but it may not exclude any agent and may involve strictly positive payments that are decreasing in allocation.

econ.TH↗

arXiv2018

Jointly learning relevant subgraph patterns and nonlinear models of their indicators

Classification and regression in which the inputs are graphs of arbitrary size and shape have been paid attention in various fields such as computational chemistry and bioinformatics. Subgraph indicators are often used as the most fundamental features, but the number of possible subgraph patterns are intractably large due to the combinatorial explosion. We propose a novel efficient algorithm to jointly learn relevant subgraph patterns and nonlinear models of their indicators. Previous methods for such joint learning of subgraph features and models are based on search for single best subgraph features with specific pruning and boosting procedures of adding their indicators one by one, which result in linear models of subgraph indicators. In contrast, the proposed approach is based on directly learning regression trees for graph inputs using a newly derived bound of the total sum of squares for data partitions by a given subgraph feature, and thus can learn nonlinear models through standard gradient boosting. An illustrative example we call the Graph-XOR problem to consider nonlinearity, numerical experiments with real datasets, and scalability comparisons to naive approaches using explicit pattern enumeration are also presented.

cs.LG↗