Jurij Leskovec
Stanford University, CA·Computer Science
Many kinds of scientific data are really networks. Molecules are atoms linked by bonds, cells interact with other cells, and brain regions are wired together.
- Fields
Computer Science
Public Health & Epidemiology
Mathematics & Statistics
- Topics
machine learning
graph neural networks
foundation models
- Grants
- 2 known grants, through Jul 2028
What they work on
- Develop multi-level graph neural networks that handle graphs whose nodes are themselves complex structures, such as molecular graphs, for scientific foundation models.
- Integrate knowledge graphs, text, images and tabular data with graph data, and add graph signal processing, fine-tuning and in-context learning so models adapt to varied scientific tasks with scarce labels.
- Develop methods to denoise and interpret imperfect multimodal real-world data, such as mobile phone and search engine logs, to capture human behavior during epidemics in real time.
- Couple agent-based disease models with cost-benefit models of behavior, and build calibration, scenario modeling and causal inference tools to forecast epidemics and estimate effects of public health decisions.
More about the research
Data structured this way are called graphs, made of points (nodes) joined by links (edges). Machine learning on graphs can help scientists predict and understand complex systems, from molecules to epidemics. These projects develop general machine learning methods for graph and multimodal data and apply them to scientific discovery and to predicting how human behavior shapes disease spread.
Why it matters
Scientists could use graph foundation models to predict properties of molecules, proteins and brain networks with little labeled data, which could speed discovery. Public health officials could use behavior-aware epidemic models to forecast outbreaks more realistically and to estimate how policies like distancing rules or vaccination campaigns affect disease spread.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Meng Jiang
University of Notre Dame, IN·Computer Science·Early-career
Develop and compare natural language generation models, one data-driven and one based on theories of abduction, for comparative summarization and argument generation in scientific domains.
Jundong Li
University of Virginia, VA·Computer Science·Early-career
Design graph neural networks tailored to peptides, including amino acid-aware message passing and graph transformers, and combine them with pretrained protein language models to merge structural and evolutionary information.
Shuai Zhang
New Jersey Institute of Technology, NJ·Engineering
Develop topology sampling and graph neural network pruning methods for single-layer graph models, trimming network structure and model size so spatial-temporal AIoT data can be analyzed more efficiently.
The school
Stanford University: 87 labs on PI Pages, $1.6B a year on research, 739 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.