252 papers
This paper surveys the theoretical and practical integration of artificial intelligence techniques, such as deep learning and reinforcement learning, into nonparametric estimation to address high-dimensional challenges and complex dependencies in fields like medicine and economics, while critically examining trade-offs between interpretability, computational cost, and statistical guarantees.
This paper demonstrates that cognitively detailed agent-based models of social influence with confirmation bias can be accurately reduced to a low-dimensional, analytically tractable mean-field system that captures the transition from consensus to polarization and provides a simple stability test for diagnosing symmetry breaking.
Using data from the 2022 and 2024 Health Information National Trends Survey, this study reveals that trust in government health agencies is the strongest predictor of perceiving climate change as a health risk, while trust in scientists shows a weaker, diminishing association over time and trust in doctors is not a significant factor.
This empirical study validates a mathematical model demonstrating that adult learners' rate of cognitive adaptation in technical fields is primarily driven by the ratio of intellectual humility to ego-noise, rather than initial intellectual potential, with minimizing ego-noise significantly accelerating the time to reach a plateau of correct solutions.
This paper proposes a scalable Random Survival Forests method that utilizes risk-set sampling to significantly reduce training time while maintaining or even improving predictive performance, with the approach now implemented in the widely used R package `ranger`.
This paper introduces a new flexible family of Pareto-type distributions generated via the right-sided Riemann–Liouville fractional integral operator, which offers a theoretically grounded mechanism for regulating tail behavior and demonstrates superior performance in modeling heavy-tailed real-world data compared to existing models.
This study evaluates large-scale propensity score adjustment strategies across four national healthcare databases and concludes that inverse probability of treatment weighting with Crump trimming and 1:1 matching generally offer the best performance, though no single strategy is universally superior, necessitating the use of diagnostics and empirical calibration to guide selection.
This study proposes an uncertainty-aware probabilistic framework that utilizes Gaussian Process Regression to effectively reconstruct sparse UAV-based radiation monitoring data from the Chornobyl Exclusion Zone, outperforming traditional interpolation methods in accuracy while providing essential spatial uncertainty estimates for informed environmental decision-making.
This paper introduces a class of robust regularised M-estimators for high-dimensional regression that achieves optimal convergence rates and variable selection consistency under minimal moment conditions (finite ), outperforming existing methods in both simulation and real-world genomic applications when dealing with heavy-tailed, skewed, or contaminated errors.
The paper proposes Sparse Convex Biclustering (SpaCoBi), a novel convex optimization method with a stability-based tuning criterion that effectively addresses noise and computational challenges in high-dimensional datasets, demonstrating superior accuracy and robustness compared to existing state-of-the-art biclustering techniques.