Mojdeh Saadati
Ph.D. Candidate in Computer Science & Applied Mathematics @ Iowa State University
Hi! I am Mojdeh. I am a Ph.D. candidate in Computer Science and Applied Mathematics at Iowa State University, advised by Prof. Baskar Ganapathysubramanian, Anson Marston Distinguished Professor of Engineering, and Prof. James Rossmanith. My doctoral work lies at the intersection of machine learning, applied mathematics, and computational modeling, with strong connections to mechanical engineering and the physical sciences.
I am interested in how machine learning can expand the limits of scientific modeling— turning computationally expensive simulations into tools for large-scale exploration, discovery, and decision making. My broader goal is to develop AI methods that do more than predict: methods that help us understand complex systems, quantify what we do not know, evaluate alternative scenarios, and uncover scientific insights that would be difficult to obtain through conventional simulation alone.
Before my Ph.D., I earned an M.S. in Computer Science from Iowa State University, advised by Prof. Jin Tian, a researcher in causal inference and probabilistic graphical models and former Ph.D. student of Judea Pearl at UCLA. My master's research focused on causal inference with missing data. I received my B.S. in Computer Science from the University of Tehran.
Research
- Scientific Machine Learning & AI for Scientific Discovery. Developing surrogate, emulator, and differentiable simulation models that transform computationally expensive mechanistic systems into scalable tools for scientific exploration, optimization, and what-if analysis.
- Uncertainty, Robustness & Distribution Shift. Quantifying predictive uncertainty and detecting out-of-distribution inputs so machine-learning systems can communicate when their predictions may be unreliable.
- Learning and Decision Making in Dynamical Systems. Developing machine-learning approaches for systems that evolve over time, with a focus on reinforcement learning for sequential decision making under uncertainty.
- Causal & Statistical Learning. Studying causal-effect identification under missing data and statistical methods for complex, incomplete observational systems.
News + Updates
Machine-readable profile: profile.json · publications.json · llms.txt