01
Latent Space Network Autoregression
A unified framework for network time series and causal peer-effect inference that embeds latent homophily into autoregressive dynamics, with asymptotic theory for finite- and infinite-time regimes.
Jae Ho Chang
Statistics · Biostatistics · Data Science
Postdoctoral researcher at Yale University developing trustworthy statistical learning under dependence, heterogeneity, and decentralization.
🔬 Fields of Interest: Mental Health, EHR
News
Updates
-
Selected for the CRSE Workshop at the Wharton School, University of Pennsylvania (16 of 50 projects) to present heterogeneous transfer learning for digital twins in quadrotor control, with NSF MATH-DT support and Aerospace Engineering at OSU.
-
Spoke at JSM 2026 in Boston in the session Transfer Learning through Empirical Risk Minimization on Personalized Federated Learning via Variance-Aware Nonparametric Empirical Bayes.
-
Presented a poster on VANEB at the IMSI workshop New Horizons on Model Transportability and Data Integration in Chicago.
-
Received the Best Student Poster Award at the ASA Georgia Chapter Workshop on Biostatistics and Bioinformatics, and the D. Ransom & Marian Whitney Award for Outstanding Research Associate at OSU Statistics.
-
Presented heterogeneous transfer learning at the OSU AI Research Summit and the Edward F. Hayes Advanced Research Forum, and spoke on Embedding Network Autoregression at the Virtual Workshop for Junior Researchers in Time Series.
-
Presented heterogeneous transfer learning at the University of Florida Winter Workshop Frontiers in Learning Under Data Heterogeneity.
-
Attended the Inaugural Workshop on Frontiers in Statistical Machine Learning at Vanderbilt, and presented the same work in a JSM topic-contributed session.
Research
Learning under structure and missing information
My work builds statistical frameworks that quantify how dependence, feature mismatch, and decentralization distort inference—and establishes principles that keep learning reliable, interpretable, and private.
02
Heterogeneous Transfer Learning
Transfer across datasets with non-overlapping feature spaces via proxy-domain mappings and discrepancy correction—with finite-sample bounds and applications in oncology survival and multi-omics integration.
03
Federated Empirical Bayes
Personalized federated learning that treats client estimators as noisy draws from an unknown prior, estimated via nonparametric maximum likelihood, with oracle denoising inequalities for privacy-preserving information exchange.
Selected papers
Recent work
-
Embedding Network Autoregression for Time Series Analysis and Causal Peer Effect Inference
Journal of Machine Learning Research (JMLR), 27(111):1–72, 2026.
-
Personalized Federated Learning via Variance-Aware Nonparametric Empirical Bayes
arXiv preprint.
-
Recidivism and Peer Influence with LLM Text Embeddings in Low Security Correctional Facilities
Revise and resubmit.
-
Heterogeneous Transfer Learning for High-Dimensional Regression with Feature Mismatch
arXiv preprint.
Path
Experience
2026–
Postdoctoral Researcher
Yale University · Faculty mentor: Prof. Yize Zhao
Exploring research questions in mental health and EHR.
2021–2026
Ph.D. in Statistics
The Ohio State University · Advisors: Prof. Subhadeep Paul and Prof. Arnab Auddy
Dissertation: Methodological Advances in Network Inference and Information Transfer.
2012–2020
B.A. & M.A. in Applied Statistics
Konkuk University, South Korea · Advisors: Prof. Kyusang Yu and Prof. Sunghoon Kwon
Recognition
Awards & honors
Selected distinctions for research excellence and support to present work across statistics, biostatistics, and machine learning venues.
May 2026
Best Student Poster Awards
Among top 5 of 20 · The 11th Workshop on Biostatistics and Bioinformatics
American Statistical Association (ASA) Georgia Chapter · Georgia State University
May 2026
D. Ransom & Marian Whitney Award for Outstanding Research Associate
Department of Statistics, The Ohio State University
Recognizing outstanding research contributions as a graduate research associate.
2025–2026
Travel Awards
Conferences and workshops across the United States
- Workshop on Collaborative Research at the Intersection of Statistics and Engineering (CRSE), Wharton School, University of Pennsylvania
- Workshop on New Horizons on Model Transportability and Data Integration, Institute for Mathematical and Statistical Innovation (IMSI)
- The 11th Workshop on Biostatistics and Bioinformatics, Georgia State University
- Annual Winter Workshop on Frontiers in Learning Under Data Heterogeneity, University of Florida
- The Inaugural Workshop on Frontiers in Statistical Machine Learning (FSML), Institute of Mathematical Statistics (IMS)
Teaching
Rigor with reach
I treat statistical thinking as a language for reasoning about evidence—helping students move from intuition to formality through visuals, simulation, and reproducible computation before asymptotics.
At OSU I led recitations in mathematical statistics and statistical inference for over a hundred students each semester, and I am prepared to teach probability, regression, computing, and modern machine learning, including electives on network time series and spectral methods.