Robust tests for log-normal lifetimes under cyclic-stress accelerated tests and interval monitoring
本文针对循环应力加速寿命测试下的对数正态寿命问题，提出基于加权最小密度幂散度估计的稳健检验方法，解决了经典检验在数据污染下显著性水平失控的问题。
0 citationsRead paper
Selected work
Representative Papers
Robust Tests for Step-Stress Models under Exponential Lifetimes
本文针对高可靠性产品寿命测试中的数据不足问题，提出基于最小密度幂散度估计器的稳健检验统计量方法来处理步进应力加速寿命试验下的指数分布寿命模型。
0 citationsRead paper
S-matrix informed neural networks for amplitude analysis
本文提出S-矩阵信息神经网络解决粒子物理中散射振幅重构问题，直接从数据学习并尊重基本原理，同时开发新数据选择程序。
0 citationsRead paper
TalentCLEF at CLEF2026: Skill and Job Title Intelligence for Human Capital Management
This study addresses key challenges in human capital management—namely, job-candidate matching, skill identification, and fairness—by introducing two core tasks: contextualized job-person matching and job-skill alignment with fine-grained skill type classification. The work presents the first NLP evaluation framework that integrates privacy preservation, multilingual adaptability, and fairness considerations, accompanied by a publicly released benchmark dataset grounded in real-world applications. By establishing this comprehensive infrastructure, the research advances the development of reproducible, cross-industry, and multilingual intelligent talent-matching systems and offers a novel paradigm for skill modeling and fairness evaluation in AI-driven human resource technologies.
0 citationsRead paper
On the suitability of ratio variables in data envelopment analysis: An application to education with methodological extensions
This study addresses the limitation of traditional data envelopment analysis (DEA) in handling ratio-type variables, which has hindered its application in international educational assessments such as PISA. The authors propose a novel DEA framework tailored for fully ratio-based inputs and outputs, establishing equivalence between the variable returns-to-scale ratio model and the constant returns-to-scale volumetric model to enable fair efficiency measurement across OECD countries. Innovatively incorporating the index of economic, social, and cultural status as an input, the framework extends both radial and directional distance functions and integrates advanced techniques—including adaptive convex envelope splines (ACES), stochastic chance constraints, and fuzzy DEA—to enhance model robustness and usability. Empirical application to PISA data facilitates equitable cross-socioeconomic comparisons of educational performance, offering a generalizable and reproducible analytical tool for international assessments.
0 citationsRead paper
Recent publications
Latest Papers
Robust tests for log-normal lifetimes under cyclic-stress accelerated tests and interval monitoring
本文针对循环应力加速寿命测试下的对数正态寿命问题，提出基于加权最小密度幂散度估计的稳健检验方法，解决了经典检验在数据污染下显著性水平失控的问题。
0 citationsRead paper
Robust Tests for Step-Stress Models under Exponential Lifetimes
本文针对高可靠性产品寿命测试中的数据不足问题，提出基于最小密度幂散度估计器的稳健检验统计量方法来处理步进应力加速寿命试验下的指数分布寿命模型。
0 citationsRead paper
S-matrix informed neural networks for amplitude analysis
本文提出S-矩阵信息神经网络解决粒子物理中散射振幅重构问题，直接从数据学习并尊重基本原理，同时开发新数据选择程序。
0 citationsRead paper
TalentCLEF at CLEF2026: Skill and Job Title Intelligence for Human Capital Management
This study addresses key challenges in human capital management—namely, job-candidate matching, skill identification, and fairness—by introducing two core tasks: contextualized job-person matching and job-skill alignment with fine-grained skill type classification. The work presents the first NLP evaluation framework that integrates privacy preservation, multilingual adaptability, and fairness considerations, accompanied by a publicly released benchmark dataset grounded in real-world applications. By establishing this comprehensive infrastructure, the research advances the development of reproducible, cross-industry, and multilingual intelligent talent-matching systems and offers a novel paradigm for skill modeling and fairness evaluation in AI-driven human resource technologies.
0 citationsRead paper
On the suitability of ratio variables in data envelopment analysis: An application to education with methodological extensions
This study addresses the limitation of traditional data envelopment analysis (DEA) in handling ratio-type variables, which has hindered its application in international educational assessments such as PISA. The authors propose a novel DEA framework tailored for fully ratio-based inputs and outputs, establishing equivalence between the variable returns-to-scale ratio model and the constant returns-to-scale volumetric model to enable fair efficiency measurement across OECD countries. Innovatively incorporating the index of economic, social, and cultural status as an input, the framework extends both radial and directional distance functions and integrates advanced techniques—including adaptive convex envelope splines (ACES), stochastic chance constraints, and fuzzy DEA—to enhance model robustness and usability. Empirical application to PISA data facilitates equitable cross-socioeconomic comparisons of educational performance, offering a generalizable and reproducible analytical tool for international assessments.
0 citationsRead paper