Guodong Chen
Postdoctoral Researcher · UC Berkeley & Lawrence Berkeley National Laboratory
Scientific machine learning · Subsurface energy systems · Computational geoscience
Research profile
I develop computational and machine-learning methods for subsurface energy and environmental systems, with particular interests in forward simulation, inverse modelling, uncertainty quantification, and optimization. My work addresses problems in geothermal energy, geological CO2 storage, fractured-media flow, and sustainable reservoir development.
A recurring objective is to connect physical modelling with modern learning methods so that complex subsurface systems can be characterized and optimized from sparse observations without sacrificing physical consistency. My current work at UC Berkeley and LBNL extends this direction toward transferable scientific models and autonomous workflows for geoscience.
Scientific machine learning
Fast and physically grounded surrogates for coupled subsurface flow and geothermal multiphysics.
Inverse problems & uncertainty
Generative inference of fractures, flow states, and geological structure from sparse observations.
Optimization & decision-making
Surrogate-assisted design and operational optimization under geological and model uncertainty.
Appointments & education
Selected publications
Multi-fidelity machine learning with knowledge transfer enhances geothermal energy system design and optimization