jlee457 [at] ur [dot] rochester [dot] edu · alyssa8921 [at] gmail [dot] com · Rochester, NY
About
I build machine learning methods that hold up when the deployment hospital, scanner,
or instrument is not the one the model was trained on.
Domain shift & OOD generalization
Continual learning
Weakly supervised learning
Foundation model adaptation
Structured prediction with LLMs
I started in computational pathology during my Ph.D. in Medicine, then entered the
Department of Brain and Cognitive Engineering at Korea University, where I completed graduate
coursework in Pattern Recognition, Computer Vision, Machine Learning, and Mathematical
Foundations of AI (GPA 4.5/4.5). Years of research in academic pathology give me access to
multi-hospital data and the judgment to evaluate what actually transfers.
I am applying to CS Ph.D. programs for Fall 2027.
I hold three separately conferred Korean degrees (M.D., M.Sc. in Medicine, Ph.D. in Medicine) and am
board-certified in pathology. “Jongwon” and “Jongwoon” are two romanizations of the same name (이종원).
News
Sep 2026Oral presentation at the MÉTIS workshop, MICCAI 2026 (Strasbourg):
Label-Free Adaptation of IHC Grading Across Hospitals.
Sep 2026Poster at CLINICCAI, MICCAI 2026 (Strasbourg).
Sep 2026Joined Prof. Sean Crowell's lab as a Graduate Research Assistant, working on
cross-instrument domain shift in satellite CO₂ retrieval.
Sep 2026Where to Decompose advanced to Phase 2 review at AAAI-27.
Sep 2026Neither Space AloneAccepted at the
ASCI Workshop, NeurIPS 2026.
Jun 2026Separating Scanner from SignalAccepted at
CLINICCAI, MICCAI 2026.
Jan 2026Started the M.S. in Computer Science at the University of Rochester.
2025First Best Poster Prize, 37th European Congress of Pathology (Vienna).
Research
Cross-Instrument Transfer in Satellite CO₂ RetrievalSep 2026 –
Lab of Prof. Sean Crowell, University of Rochester
Domain shift between instruments in remote-sensing retrieval of atmospheric CO₂.
3D Meissner Corpuscle Detection in In Vivo Confocal MicroscopySep 2026 –
Dept. of Neurology (Prof. D. Herrmann) and Lab of Prof. Jiebo Luo, Dept. of Computer Science, University of Rochester
Detecting nerve endings across serial reflectance confocal sections of human skin: hard-negative
fine-tuning of a YOLO detector and cross-slice association of 2D detections into 3D objects.
Structured Prediction with Language Models for Electrodiagnostic InterpretationMay 2026 –
Dept. of Neurology and Lab of Prof. Jiebo Luo, Dept. of Computer Science, University of Rochester
A two-stage pipeline that turns free-text EMG/NCS studies into structured diagnostic slots and back into
a natural-language report, composing six domain-specialist models through a rule-based arbiter.
Diagnostic-level F1 0.89 on 404 held-out studies, against 0.84 for a single fine-tuned baseline.
Cross-Hospital Adaptation of Pathology Foundation Models2025 –
University of Rochester / Korea University Guro Hospital
Label-free adaptation of frozen UNI and Virchow2 features across three hospital cohorts for
Nectin-4, HER2 and TROP2 immunohistochemistry grading, with input-space and feature-space corrections.
Weakly Supervised Prognostic Modeling in Digital Pathology2022 –
Asan Medical Center / Korea University Guro Hospital
Attention-based multiple-instance learning on whole-slide images for recurrence and survival prediction
in bladder, colorectal, oral tongue and bile duct carcinoma; AI-based tumor-infiltrating lymphocyte quantification.
Publications
* first author † corresponding author
Conference and Workshop Papers
Separating Scanner from Signal: FFT-Based Frequency Adaptation for Cross-Hospital IHC Expression Grading
Neither Space Alone: Input- and Feature-Space Correction for Cross-Hospital IHC Grading
Jongwon Lee*, Jhair Gallardo
AI at Scale for Clinical Impact (ASCI) Workshop, NeurIPS 2026Accepted
Where to Decompose: Expert Composition and Structured Prediction for Electrodiagnostic Interpretation
Jongwon Lee*
AAAI Conference on Artificial Intelligence (AAAI-27)Under review · Phase 2
Journal Articles
Field-of-View Impact on Multi-Class Tissue Segmentation: A Systematic Analysis of Size-Dependent Detection Performance Across Multiple Cancer Types
D Jo, Jongwon Lee*, et al.
Laboratory Investigation, 2026 Minor revision
Artificial Intelligence-Assessed Tumor Central Tumor-Infiltrating Lymphocytes Predict Recurrence-Free Survival in Transurethral Resection of Bladder Tumor Specimens
S Lee, Jongwon Lee†, et al.
Laboratory Investigation, 2026 Major revision
The Expression and Clinical Relevance of LRRC15+ Cancer-Associated Fibroblasts in Colorectal Cancer
Y Yang, D Jung, Jongwon Lee*, et al.
Journal of Gastroenterology and Hepatology, 2026 Major revision
Assessment of Established Prognostic Factors and Artificial Intelligence-Based Evaluation of Tumor-Infiltrating Lymphocytes in Oral Tongue Squamous Cell Carcinoma
Jongwon Lee*, Ming Fan, et al.
Oral Oncology, 2025 · discussed in an invited commentary (Oral Oncology 2025;168:107586)
AI Prediction of Clinicopathological Factors and LRRC15-Immunolabeled Cancer-Associated Fibroblast Level in Colorectal Carcinoma Using Attention-Based Multiple Instance Learning
37th European Congress of Pathology 2025, Vienna · First Best Poster Prize