Talks

Talks and slides

Presentations with files to view or download; most are tied to my papers. Conference talks are listed on the CV.

JHU Biostatistics Retreat Presentation, October 2, 2026

Causal inference with a right-censored marker of disease progression

How to adjust for disease severity when it is not measured. A proxy built from the age at disease onset and the current age carries similar information, but the age at onset is right-censored. The page sets up the statistical problem and shows simulation results for the complete case and IPW estimators across censoring rates.

Open PDF Related paper

COMPSTAT 2026, Athens, Greece, August 27, 2026

Matching estimators to censoring rate improves inference with outcome-dependent right-censored covariates

Talk on Huntington disease data where the age at onset is right-censored and censoring depends on the outcome. The talk compares estimators across censoring rates and shows that choosing the estimator to match the censoring rate improves inference.

Paper on arXiv

International Biometric Conference, Seoul, South Korea, July 14, 2026

Federated learning with incomplete data: when to use complete cases and when to weight

Talk on federated learning when each site has missing data, motivated by pleural infection outcomes across hospitals. The talk covers when the complete case estimator is enough and when a weighted estimator is needed, and how to correct its variance.

Paper on arXiv

CIDI T32 Meeting, Johns Hopkins, March 10, 2026

Federated learning with incomplete data: a weighted approach with variance correction

A lecture for the CIDI T32 trainees on federated learning when sites have missing data. The slides introduce distributed research networks and the COLA-GLM estimator, which assumes complete data, and then develop a weighted version for missing covariates, with simulation results on bias, coverage, and the variance correction. The example comes from pleural infection data at two sources.

Paper on arXiv

Stuart Lab Meeting, Johns Hopkins Biostatistics, November 4, 2025

How we work together: lab values

Slides on the values that guide how the lab works together: a mindset of improving the lab, engaging in lab meetings and operations, sharing skills through workshops and co-working sessions, and building community by supporting each other. The slides close with a prompt for each member to propose a team agreement.