On the disintegration of the stochastic majority vote: From PAC-Bayesian bounds to a self-bounding algorithm
本文解决了随机多数投票的确定性转化问题，通过应用最新的解构PAC-Bayesian理论直接到权重向量空间，提出了一个自我约束的学习算法。
0 citationsRead paper
Selected work
Representative Papers
Regional advantage in rugby sevens: Is there a home effect when nobody is home?
This study investigates regional advantage in rugby sevens within contexts lacking formal home venues. Analyzing 2,672 matches, the research innovatively replaces traditional binary home-away variables with continuous measures of geographic, spatiotemporal, and cultural proximity within a symmetric team-match panel fixed-effects model. Results indicate that while aggregate regional advantage is statistically insignificant, substantial heterogeneity exists; specifically, certain teams exhibit performance declines correlated with increased travel distance and east-west time zone differences. These findings reveal the team-specific nature of regional advantages, offering novel empirical evidence and theoretical perspectives for understanding competitive performance in non-traditional home environments. This work thereby advances the literature on spatial factors in sports analytics by demonstrating how nuanced proximity metrics can capture complex performance dynamics absent in conventional venue-based frameworks.
0 citationsRead paper
The Best Are Always the Best: COVID-19 Lockdown Stringency and the Dispersion of Olympic Medal Outcomes
This study investigates the impact of COVID-19 lockdown stringency on national Olympic performance and medal distribution. Analyzing data from 99 countries alongside the Oxford Stringency Index through OLS regression and ANOVA, we find that while lockdowns did not alter global average competitive standards, they significantly exacerbated medal dispersion in men’s events among non-traditional sporting powers. Specifically, high-stringency conditions increased this dispersion three- to six-fold. These findings reveal that external shocks preserve established competitive hierarchies while amplifying performance uncertainty for disadvantaged teams. Consequently, this work provides novel evidence regarding the heterogeneous effects of public health emergencies on equity in elite sports competition, highlighting how systemic disruptions disproportionately affect marginal participants without necessarily diminishing overall global performance levels.
0 citationsRead paper
Non-obvious Manipulability with Groups in Shapley-Scarf Housing Markets
This study addresses the mechanism design challenge in Shapley-Scarf housing markets arising from the relaxation of strategy-proofness. We propose group non-explicit manipulability and group rationality conditions to construct a class of super-group Top Trading Cycles (TTC) mechanisms. Leveraging game theory and matching theory, this work effectively expands the design space for non-strategy-proof mechanisms while preserving Pareto efficiency. The proposed mechanism class simultaneously satisfies group rationality and efficiency requirements, achieving consistency between best and worst-case outcomes. Consequently, this research establishes a novel paradigm for housing market mechanism design that integrates theoretical innovation with practical applicability, offering robust solutions where strict strategy-proofness is relaxed.
0 citationsRead paper
On the post-hoc Evaluation of PDE Discovery: A Multifaceted Challenge of Scientific Advancement
Current evaluation practices for partial differential equation (PDE) discovery lack a unified standard, and existing metrics are often narrow in scope, failing to simultaneously account for predictive accuracy, physical consistency, interpretability, and out-of-distribution generalization—potentially leading to erroneous identification of novel physical laws. This work establishes the first systematic classification framework for post-discovery PDE evaluation, integrating techniques from machine learning, numerical analysis, information theory, and symbolic regression to holistically assess model performance across multiple dimensions, including prediction fidelity, adherence to physical constraints, model simplicity, and generalization capability. By exposing the limitations of prevailing evaluation approaches, this study proposes a standardized and extensible evaluation paradigm that provides both algorithm developers and scientific practitioners with a rigorous methodological foundation for reliably validating newly discovered physical laws.
0 citationsRead paper
Recent publications
Latest Papers
On the disintegration of the stochastic majority vote: From PAC-Bayesian bounds to a self-bounding algorithm
本文解决了随机多数投票的确定性转化问题，通过应用最新的解构PAC-Bayesian理论直接到权重向量空间，提出了一个自我约束的学习算法。
0 citationsRead paper
Regional advantage in rugby sevens: Is there a home effect when nobody is home?
This study investigates regional advantage in rugby sevens within contexts lacking formal home venues. Analyzing 2,672 matches, the research innovatively replaces traditional binary home-away variables with continuous measures of geographic, spatiotemporal, and cultural proximity within a symmetric team-match panel fixed-effects model. Results indicate that while aggregate regional advantage is statistically insignificant, substantial heterogeneity exists; specifically, certain teams exhibit performance declines correlated with increased travel distance and east-west time zone differences. These findings reveal the team-specific nature of regional advantages, offering novel empirical evidence and theoretical perspectives for understanding competitive performance in non-traditional home environments. This work thereby advances the literature on spatial factors in sports analytics by demonstrating how nuanced proximity metrics can capture complex performance dynamics absent in conventional venue-based frameworks.
0 citationsRead paper
The Best Are Always the Best: COVID-19 Lockdown Stringency and the Dispersion of Olympic Medal Outcomes
This study investigates the impact of COVID-19 lockdown stringency on national Olympic performance and medal distribution. Analyzing data from 99 countries alongside the Oxford Stringency Index through OLS regression and ANOVA, we find that while lockdowns did not alter global average competitive standards, they significantly exacerbated medal dispersion in men’s events among non-traditional sporting powers. Specifically, high-stringency conditions increased this dispersion three- to six-fold. These findings reveal that external shocks preserve established competitive hierarchies while amplifying performance uncertainty for disadvantaged teams. Consequently, this work provides novel evidence regarding the heterogeneous effects of public health emergencies on equity in elite sports competition, highlighting how systemic disruptions disproportionately affect marginal participants without necessarily diminishing overall global performance levels.
0 citationsRead paper
Non-obvious Manipulability with Groups in Shapley-Scarf Housing Markets
This study addresses the mechanism design challenge in Shapley-Scarf housing markets arising from the relaxation of strategy-proofness. We propose group non-explicit manipulability and group rationality conditions to construct a class of super-group Top Trading Cycles (TTC) mechanisms. Leveraging game theory and matching theory, this work effectively expands the design space for non-strategy-proof mechanisms while preserving Pareto efficiency. The proposed mechanism class simultaneously satisfies group rationality and efficiency requirements, achieving consistency between best and worst-case outcomes. Consequently, this research establishes a novel paradigm for housing market mechanism design that integrates theoretical innovation with practical applicability, offering robust solutions where strict strategy-proofness is relaxed.
0 citationsRead paper
On the post-hoc Evaluation of PDE Discovery: A Multifaceted Challenge of Scientific Advancement
Current evaluation practices for partial differential equation (PDE) discovery lack a unified standard, and existing metrics are often narrow in scope, failing to simultaneously account for predictive accuracy, physical consistency, interpretability, and out-of-distribution generalization—potentially leading to erroneous identification of novel physical laws. This work establishes the first systematic classification framework for post-discovery PDE evaluation, integrating techniques from machine learning, numerical analysis, information theory, and symbolic regression to holistically assess model performance across multiple dimensions, including prediction fidelity, adherence to physical constraints, model simplicity, and generalization capability. By exposing the limitations of prevailing evaluation approaches, this study proposes a standardized and extensible evaluation paradigm that provides both algorithm developers and scientific practitioners with a rigorous methodological foundation for reliably validating newly discovered physical laws.
0 citationsRead paper