Scientific machine learning is a growing field that sits at the intersection of physics and data: By combining physics-based models with machine learning — a form of artificial intelligence that designs computer algorithms to detect patterns in data and use these patterns to make predictions about new data — researchers are developing new ways to simulate, predict and understand complex, real-world systems from aquifers to living tissue.
A dedicated cluster of researchers at Rice University’s Ken Kennedy Institute is using scientific machine learning to explore innovative solutions to complex science and engineering problems. The research cluster focuses on fundamental research while developing algorithms and theories with applications in a wide range of contexts such as geoscience, computational biology, medicine and the energy sector.
The following Rice experts can answer questions about emerging trends, challenges and opportunities in scientific machine learning:
- Beatrice Riviere develops computational methods to model physical phenomena characterized by conservation laws. Her work in porous materials supports applications in energy systems, carbon storage, environmental modeling and computational biomedicine.
- Matthias Heinkenschloss develops mathematical and computational tools for solving complex decision-making and design problems in science and engineering with applications in areas such as trajectory optimization, parameter identification in neuron systems, reservoir management and the optimization of other systems with a large parameter space.
- Lu Zhang designs and analyzes mathematical models of physical and biological systems, including wave propagation, seismic imaging and population dynamics. Her work emphasizes structure-preserving numerical and data-driven methods that combine scientific machine learning with rigorous mathematical analysis to develop reliable, interpretable models for real-world applications.
Highlighting the growing relevance of scientific machine learning for energy applications, the group is organizing a workshop at the 2026 Energy HPC & AI Conference that will convene researchers from industry, national laboratories and universities to discuss advances in scientific machine learning (SciML) in engineering applications.
“This is an exciting time to conduct research at the forefront of scientific machine learning,” said Riviere, who is the research cluster lead and Rice’s Noah Harding Chair and Professor of Computational Applied Mathematics and Operations Research. “By integrating physical laws with data-driven models, our cluster of researchers develops new, efficient, robust, interpretable and data-efficient computational tools to advance simulation, inference and decision-making in complex applications. Our SciML workshop brings together a group of experts who will highlight the latest advances in the field, including generative AI for design optimization, sensitivity-based model inference and operator learning for engineering applications.”
View details about the SciML workshop by visiting the conference schedule.