Yeonju Lee
Hello! I’m a Ph.D. student in the H. Milton Stewart School of Industrial & Systems Engineering at Georgia Tech, advised by Dr. Jing Li.
My research focuses on knowledge-informed machine learning (ML) , developing models that integrate domain knowledge to build more data-efficient, robust, and interpretable systems. Please feel free to contact me at ylee845@gatech.edu.
CV (Updated Oct 2025) | Google Scholar
Knowledge-Informed ML: Why, Where, and How
Why Knowledge-Informed ML?
Real-world data often suffers from scarcity, heterogeneity, and high dimensionality, making it difficult for purely data-driven models to generalize. Knowledge-informed ML integrates domain understanding to enhance robustness and interpretability.
Where to Inform?
At the input, architecture, and inference levels — where domain knowledge can guide representation, structure, and decision-making.
How to Inform?
Through representations, constraints, and rules — translating expert understanding into model priors and learning objectives.
Applications
- Healthcare – dental lesion detection and medical image translation
- Precision Agriculture – yield forecasting, anomaly detection, and change-point detection
- Manufacturing – root cause analysis and process optimization
For more details, please refer to the Publications section.