Multivariate Spatio-Temporal Regression with Penalized Model Selection and an Empirical Application
本文提出一种多变量时空回归框架，通过惩罚模型选择方法解决空间、时间和跨方程依赖问题，并在日本关西地区社会经济数据中验证了其有效性。
0 citationsRead paper
Selected work
Representative Papers
Does medical specialization of VLMs enhance discriminative power?: A comprehensive investigation through feature distribution analysis
Current evaluation protocols struggle to determine whether medical vision-language models (VLMs) genuinely learn lesion-specific discriminative features. This work presents the first systematic comparison—through feature distribution analysis—of representation capabilities between domain-specific and general-purpose VLMs on multimodal medical images, employing both visualization and quantitative assessment across multiple lesion classification datasets. The study reveals that contextual enhancement in the text encoder is more critical for improving discriminability than large-scale medical image pretraining; that enhanced non-medical VLMs (e.g., LLM2CLIP) can outperform specialized models; and that general-purpose VLMs are susceptible to biases introduced by overlaid text in medical images. These findings offer a novel perspective for the design and evaluation of medical VLMs.
0 citationsRead paper
Risk Sharing Among Many: Implementing a Subgame Perfect and Optimal Equilibrium
This paper addresses the problem of ex ante welfare-maximizing risk allocation among a large population of agents, extending Echenique and Núñez’s (2025) “Prices and Choices” (P&C) mechanism to infinite-dimensional risk allocation spaces. While the original P&C mechanism is restricted to finite choice sets, we establish—under an infinite choice set defined by Lipschitz-continuous price functions—the first rigorous proof that the mechanism still induces a subgame-perfect Nash equilibrium and achieves global utilitarian welfare maximization. Methodologically, the analysis integrates tools from game theory, infinite-dimensional mechanism design, and Lipschitz function theory. The key contribution is the theoretical expansion of the P&C mechanism’s domain of applicability: we provide the first implementable, equilibrium-guaranteed, welfare-optimal framework for continuous risk sharing, thereby overcoming a fundamental limitation of prior discrete-choice mechanisms.
0 citationsRead paper
Recent publications
Latest Papers
Multivariate Spatio-Temporal Regression with Penalized Model Selection and an Empirical Application
本文提出一种多变量时空回归框架，通过惩罚模型选择方法解决空间、时间和跨方程依赖问题，并在日本关西地区社会经济数据中验证了其有效性。
0 citationsRead paper
Does medical specialization of VLMs enhance discriminative power?: A comprehensive investigation through feature distribution analysis
Current evaluation protocols struggle to determine whether medical vision-language models (VLMs) genuinely learn lesion-specific discriminative features. This work presents the first systematic comparison—through feature distribution analysis—of representation capabilities between domain-specific and general-purpose VLMs on multimodal medical images, employing both visualization and quantitative assessment across multiple lesion classification datasets. The study reveals that contextual enhancement in the text encoder is more critical for improving discriminability than large-scale medical image pretraining; that enhanced non-medical VLMs (e.g., LLM2CLIP) can outperform specialized models; and that general-purpose VLMs are susceptible to biases introduced by overlaid text in medical images. These findings offer a novel perspective for the design and evaluation of medical VLMs.
0 citationsRead paper
Risk Sharing Among Many: Implementing a Subgame Perfect and Optimal Equilibrium
This paper addresses the problem of ex ante welfare-maximizing risk allocation among a large population of agents, extending Echenique and Núñez’s (2025) “Prices and Choices” (P&C) mechanism to infinite-dimensional risk allocation spaces. While the original P&C mechanism is restricted to finite choice sets, we establish—under an infinite choice set defined by Lipschitz-continuous price functions—the first rigorous proof that the mechanism still induces a subgame-perfect Nash equilibrium and achieves global utilitarian welfare maximization. Methodologically, the analysis integrates tools from game theory, infinite-dimensional mechanism design, and Lipschitz function theory. The key contribution is the theoretical expansion of the P&C mechanism’s domain of applicability: we provide the first implementable, equilibrium-guaranteed, welfare-optimal framework for continuous risk sharing, thereby overcoming a fundamental limitation of prior discrete-choice mechanisms.
0 citationsRead paper