About Me
I am a PhD candidate in Computing and Data Sciences at Boston University (BU), working with Jonathan Huggins and Michael Dietze. I am broadly interested in uncertainty quantification, probabilistic machine learning, and spatiotemporal modeling, with the goal of developing new methodologies for scientific applications. My current work is motivated by the problem of calibrating large-scale mechanistic models of the terrestrial carbon cycle.
Beyond my research, I am a trainee in BU URBAN, a graduate program in urban biogeoscience and environmental health. I am also an officer in the BU chapter of the American Statistical Association, and previously co-organized the Environment and Climate working group within the Mechanism Design for Social Good (MD4SG) research intitiative. In the summer of 2025, I had the pleasure of interning with the Jet Propulsion Laboratory’s Uncertainty Quantification and Statistical Analysis group.
Prior to BU, I worked as a research assistant at the Federal Reserve Bank of Boston. I received my B.S. in Quantitative Economics from Tufts University in 2018.
Contact: arober [at] bu [dot] edu
Publications and Preprints
Propagating Surrogate Uncertainty in Bayesian Inverse Problems
Andrew Gerard Roberts, Michael Dietze, and Jonathan Huggins
arXiv preprint, 2026
ArXiv Code
Parameter estimation in land surface models: Challenges and opportunities with data assimilation and machine learning
Raoult, N., Douglas, N., MacBean, N., Kolassa, J., Quaife, T., Roberts, A. G., et al.
Journal of Advances in Modeling Earth Systems, 2025
Journal