AI for Precision Medicine: Histopathology-centered Computational analysis of Spatial Omics: Integration, Mapping, and Foundation Models
Knowledge Discovery in Databases (KDD) Lecture-Style Tutorial 2026
August 9th, 2026
Jeju, Korea
Embedded Files
Abstract
Spatial omics (SO) technologies enable spatially resolved molecular profiling, while hematoxylin and eosin (H&E) imaging remains the gold standard for morphological assessment in clinical pathology. Recent computational advances increasingly place H&E images at the center of SO analysis, bridging morphology with transcriptomic, proteomic, and other spatial molecular modalities. This lecture-style tutorial surveys the algorithmic foundations and recent advances in method development that make them practical and impactful for precision medicine. Following the tutorial flow, we first introduce key SO modalities and data abstractions (tiles/patches, spots, cells, and spatial graphs) and articulate problems to address and motivations, emphasizing multi-scale mismatch, structured spatial dependence, weak supervision, and domain shift across cohorts and sites. We then trace the evolution of modern multimodal representation learning, highlighting graph neural networks, transformer-based architectures, and encoder–decoder designs. The core of the tutorial systematically organizes contemporary methods into three categories: (i) integration methods, which jointly model paired multimodal measurements; (ii) mapping methods, which predict spatial molecular profiles from H&E images; and (iii) foundation models (FMs), which learn transferable representations from large-scale spatial datasets via self-supervised pretraining, contrastive objectives, etc. This tutorial also discusses applications of generative modeling to support imputation and data augmentation. Throughout, we connect methods to real biomedical endpoints (e.g., tumor microenvironment characterization, biomarker discovery, and cohort-level stratification). We further summarize actionable modeling directions enabled by current architectures and delineate persistent gaps driven by data, biology, and technology that are unlikely to be resolved by model design alone. The tutorial concludes with open challenges in interpretability, reliability, privacy, and clinical translation, outlining opportunities for the KDD community to contribute principled data mining and learning approaches to multimodal spatial biology.
Materials
Schedule
Time: Sunday, August 9th, 2026, 8am-12pm
Location: Jeju, Korea
Outline
Background
Why is this a crucial data mining problem? What are data modalities?
Overview of spatial omics
Histopathologic H&E imaging as the anchor
Evolution of algorithms and architecture design
Significance
Spatial Data Modalities: H&E Images and Omics Data
H&E data availability and challenges
Spatial omics data and challenges
Spot-level and (sub)cell-level
Key public datasets
Current State of Art and Motivation
Statistical deconvolution of mixed signals
Deep learning for H&E-to-gene prediction, cross-modal alignment, and cell neighborhood modeling
Limitations
Motivation
Multimodal Integrative Methodology Trend
Integration
Mapping
Foundation Models
Downstream Tasks and Actionable Modeling Improvements
Downstream tasks
Spatial architecture identification
Spatial pattern recognition
Molecular profile imputation
Cell-cell network inference
Biomedical endpoint prediction
Common failures and actionable directions
Resolution-related failures
Generalization-related failures, including scale generalization, batch and technical generalization, organ and species generalization
Interpretation and evaluation-related failures
Challenges and Future Directions
Challenges and current practical limitations
Representation Gap: selection and representation of regions of interest (ROIs)
Dimensionality Gap: 2-dimensional (2D) sections to 3D/4D tissue dynamics
Technology Gap: assay harmonization, resolution gap between H&E images and omics
Future directions:
Multi-omics spatial assays (epigenomics, proteomics)
Scalable multimodal learning frameworks
Summary and Q&A
Tutors' Bio
Ninghui Hao received her master's degree of Biomedical Informatics at Harvard Medical School. Her research focuses on AI for Science (AI4Science), with interests in bioinformatics, explainable machine learning, multimodal learning, and interpretable models for clinical decision-making. Her work has been published in venues including the Association for the Advancement of Artificial Intelligence (AAAI), the American Association for Cancer Research (AACR) and others.
Boshen Yan is a Ph.D. student in Computational Biology at Carnegie Mellon University. He completed his master’s degree in Biomedical Informatics at Harvard Medical School and his bachelor’s degree in Computational Biology at the National University of Singapore. His research interests include the development of multimodal integration methods to analyze genetic diseases.
Dong Li is currently a Ph.D. student in the Department of Computer Science at Baylor University. His main research directions include graph mining, fairness-aware machine learning, domain generalization, and computational biology. He has received multiple academic scholarships and national competition awards. His publications have been accepted by top international conferences such as KDD, IJCAI, CIKM, etc.
Xintao Wu is a Professor and the Charles D. Morgan/Acxiom Endowed Graduate Research Chair and leads the Social Awareness and Intelligent Learning (SAIL) Lab in the Electrical Engineering and Computer Science Department at the University of Arkansas. Dr. Wu is an associate editor or editorial board member of several journals and program committees as area chair, senior PC, and PC of top international conferences. He has served as the program co-chair of the ACM EAI-KDD workshops (2022-2026), IEEE BigData'2022, IEEE ICMLA'2024, and PAKDD'2025. He also gave multiple tutorials on causality-based ethical AI and fair machine learning under distribution shift at top international conferences, including ACM KDD, IEEE BigData, IJCAI, and AAAI.
Chen Zhao is an Assistant Professor in the Department of Computer Science at Baylor University. His research focuses on machine learning, data mining, and computational biology, particularly trustworthy machine learning, novelty detection, and domain generalization. His publications have been accepted and published in premier conferences, including KDD, CVPR, IJCAI, ICDE, AAAI, WWW, etc. Dr. Zhao served as a PC member of top international conferences, such as KDD, NeurIPS, IJCAI, ICML, AAAI, ICLR, etc. He has organized and chaired multiple workshops on topics of Ethical AI, Uncertainty Quantification, Distribution Shifts, and Trustworthy AI for Healthcare at KDD (2022, 2023, 2024, 2025), AAAI (2023), IEEE BigData (2024), and SDM (2025). He serves as the chair of the Challenge Cup of the IEEE Bigdata 2024 conference, the tutorial chair for the PAKDD 2025 and ICDM 2025 conferences, and the workshop chair for the IEEE Bigdata 2025 conference.
Guihong Wan is an Assistant Professor in the Institute for Population and Precision Health at the University of Chicago. Her research focuses on developing computational AI methodologies for integrative analyses of biomedical data and building biologically explainable machine learning models for predicting patient outcomes. Her research has been published in AAAI, IJCAI, ICDE, The Lancet Oncology, npj Precision Oncology, Briefings in Bioinformatics, Journal of the American Academy of Dermatology, British Journal of Dermatology, Nature Medicine, and others. She served as the tutorial co-chair for the 16th International Conference on Brain Informatics, 2023.
Page updated
Google Sites
Report abuse