Research
Research Interests
My research can be categorized into the following two areas:
Learning physical models from the dynamical data sets, which involves learning interaction kernels for collective dynamics, nonparametric estimation of composite functions, and building generalizable and interpretable predictive models.
Providing performance guarantees for predictive models built with unknowns (aleatoric and epistemic uncertainties) via uncertainty quantification and sensitivity analysis, and identifying the most influential components on predictions for the probabilistic graphical models.
Publications
Data-driven model selections of second-order particle dynamics via integrating Gaussian processes with low-dimensional interacting structures,(with Kulick, C., and Tang, S.), to appear in Physica D: Nonlinear Phenomena (2024).
Learning particle swarming models from data with Gaussian processes, (with Kulick, C., Tang, S., and Ren, Y.), to appear in Mathematics of Computation (2023).
Learning Interaction Variables and Kernels from Observations of Agent-Based Systems, (with Maggioni, M., Martin, P., and Zhong, M.), IFAC-PapersOnLine 55, no. 30 (2022): 162-167.
Model Uncertainty and Correctability for Directed Graphical Models (with Birmpa, P., Rey-Bellet, L., and Katsoulakis, M. A.), SIAM/ASA Journal on Uncertainty Quantification 10, no. 4 (2022): 1461-1512.
Explainable and trustworthy artificial intelligence for correctable modeling in chemical sciences (with Lansford, J., Vlachos, D. G., and Katsoulakis, M. A.), Science advances 6.42 (2020): eabc3204.
Non-parametric correlative uncertainty quantification and sensitivity analysis: Application to a Langmuir bimolecular adsorption model (with Lansford, J., Mironenko, A., Pourkargar, D. B., Vlachos, D. G., and Katsoulakis, M. A.), AIP Advances 8.3 (2018): 035021.