Selected work

Representative Papers

Robust and Reliable AI for Predictive Quality in Semiconductor Materials Manufacturing with MLOps and Uncertainty Quantification

May 08, 2026

This study addresses the degradation of AI model performance in semiconductor manufacturing caused by process variations, equipment aging, and raw material shifts. Leveraging five years of real production line data, the work systematically evaluates multiple MLOps retraining strategies for predictive quality and integrates conformal prediction to deliver statistically valid uncertainty quantification. The authors propose an efficient fixed-interval retraining strategy—updating the model every five lots without hyperparameter tuning—that maintains high prediction accuracy under both abrupt process shifts and gradual equipment degradation while substantially reducing computational overhead. By combining normalized residual control limits with conformal prediction intervals, the approach transitions quality assurance from reactive inspection to proactive, reliable forecasting, offering a robust and practical solution for industrial AI deployment.

0 citationsRead paper

Statistical Methodology Groups in the Pharmaceutical Industry

Mar 11, 2026

This work addresses the inefficiencies in clinical trial design and analysis that hinder drug development success rates and timelines. It proposes establishing a dedicated statistical methodology team within pharmaceutical companies as a strategic investment, embedded through a systematic organizational structure and cross-functional collaboration mechanisms—both internally across departments and externally with academic and regulatory partners—to break down information silos. By integrating advanced statistical modeling, optimized clinical trial designs, and other high-impact quantitative methodologies, this team significantly enhances R&D efficiency, shortens development cycles, and strengthens the scientific rigor and likelihood of success in clinical decision-making.

0 citationsRead paper

Incorporating estimands into meta-analyses of clinical trials

Oct 17, 2025

This study addresses quantitative heterogeneity and limited external validity in clinical trial meta-analyses arising from divergent handling of intercurrent events (e.g., treatment discontinuation, rescue medication). It pioneers the systematic integration of the estimand framework into meta-analysis methodology. Methodologically, it establishes a novel estimand-based meta-analytic paradigm that distinguishes between treatment-policy and hypothetical estimands, and incorporates network meta-analysis. Using GLP-1 receptor agonists for type 2 diabetes as an empirical case study, it systematically contrasts results with those derived under the conventional PICO framework to identify sources of divergence. Key contributions include: (1) formalizing an estimand-driven evidence synthesis pathway that enhances traceability of heterogeneity and transparency of effect estimation; and (2) substantially improving the relevance and applicability of pooled estimates for health technology assessment and regulatory decision-making—thereby providing an original methodological advancement for evidence-based medicine.

0 citationsRead paper

Deep Learning-based Prediction of Clinical Trial Enrollment with Uncertainty Estimates

Jul 31, 2025

Predicting patient enrollment in clinical trials is a critical challenge for trial planning. This paper proposes a deep probabilistic forecasting model that jointly leverages unstructured protocol text and structured site-level features. It employs a pretrained language model to encode trial protocols and constructs tabular representations from site characteristics; cross-modal attention aligns textual and tabular embeddings. A Gamma-distributed output layer explicitly models the Poisson–Gamma enrollment process, enabling principled uncertainty quantification and prediction interval estimation. Evaluated on real-world multicenter clinical trial data, the model significantly outperforms conventional statistical methods and end-to-end deep learning baselines—achieving superior accuracy and reliability in both enrollment count and timeline predictions. The approach delivers interpretable, robust decision support for trial resource allocation and operational planning.

0 citationsRead paper

Recent publications

Latest Papers

Robust and Reliable AI for Predictive Quality in Semiconductor Materials Manufacturing with MLOps and Uncertainty Quantification

May 08, 2026

This study addresses the degradation of AI model performance in semiconductor manufacturing caused by process variations, equipment aging, and raw material shifts. Leveraging five years of real production line data, the work systematically evaluates multiple MLOps retraining strategies for predictive quality and integrates conformal prediction to deliver statistically valid uncertainty quantification. The authors propose an efficient fixed-interval retraining strategy—updating the model every five lots without hyperparameter tuning—that maintains high prediction accuracy under both abrupt process shifts and gradual equipment degradation while substantially reducing computational overhead. By combining normalized residual control limits with conformal prediction intervals, the approach transitions quality assurance from reactive inspection to proactive, reliable forecasting, offering a robust and practical solution for industrial AI deployment.

0 citationsRead paper

Statistical Methodology Groups in the Pharmaceutical Industry

Mar 11, 2026

This work addresses the inefficiencies in clinical trial design and analysis that hinder drug development success rates and timelines. It proposes establishing a dedicated statistical methodology team within pharmaceutical companies as a strategic investment, embedded through a systematic organizational structure and cross-functional collaboration mechanisms—both internally across departments and externally with academic and regulatory partners—to break down information silos. By integrating advanced statistical modeling, optimized clinical trial designs, and other high-impact quantitative methodologies, this team significantly enhances R&D efficiency, shortens development cycles, and strengthens the scientific rigor and likelihood of success in clinical decision-making.

0 citationsRead paper

Incorporating estimands into meta-analyses of clinical trials

Oct 17, 2025

This study addresses quantitative heterogeneity and limited external validity in clinical trial meta-analyses arising from divergent handling of intercurrent events (e.g., treatment discontinuation, rescue medication). It pioneers the systematic integration of the estimand framework into meta-analysis methodology. Methodologically, it establishes a novel estimand-based meta-analytic paradigm that distinguishes between treatment-policy and hypothetical estimands, and incorporates network meta-analysis. Using GLP-1 receptor agonists for type 2 diabetes as an empirical case study, it systematically contrasts results with those derived under the conventional PICO framework to identify sources of divergence. Key contributions include: (1) formalizing an estimand-driven evidence synthesis pathway that enhances traceability of heterogeneity and transparency of effect estimation; and (2) substantially improving the relevance and applicability of pooled estimates for health technology assessment and regulatory decision-making—thereby providing an original methodological advancement for evidence-based medicine.

0 citationsRead paper

Deep Learning-based Prediction of Clinical Trial Enrollment with Uncertainty Estimates

Jul 31, 2025

Predicting patient enrollment in clinical trials is a critical challenge for trial planning. This paper proposes a deep probabilistic forecasting model that jointly leverages unstructured protocol text and structured site-level features. It employs a pretrained language model to encode trial protocols and constructs tabular representations from site characteristics; cross-modal attention aligns textual and tabular embeddings. A Gamma-distributed output layer explicitly models the Poisson–Gamma enrollment process, enabling principled uncertainty quantification and prediction interval estimation. Evaluated on real-world multicenter clinical trial data, the model significantly outperforms conventional statistical methods and end-to-end deep learning baselines—achieving superior accuracy and reliability in both enrollment count and timeline predictions. The approach delivers interpretable, robust decision support for trial resource allocation and operational planning.

0 citationsRead paper