Tenavi Nakamura-Zimmerer

Ph.D. Candidate, Applied Mathematics and Statistics

University of California Santa Cruz

About me

I work at the exciting interface between machine learning and dynamical systems. Specifically, I seek novel ways to apply machine learning tools for modeling and computational optimal control of high-dimensional nonlinear systems under uncertainty. In addition, my research also explores how domain knowledge from physics and control theory can be used to improve such machine learning methods. More generally, I am interested in computational optimal control, dynamical systems, machine learning, and aerospace applications.

I received my B.A. from UC Santa Cruz with a double major in pure mathematics and language studies in 2015. Since 2016, I have been working towards a Ph.D. in applied mathematics and statistics, also at UC Santa Cruz. In 2020 I started a Pathways Internshep at NASA, first at Ames Research Center in the Aviation Systems division and now at Langley Research Center in the Flight Dynamics branch.

My CV

Contact

Email: tenakamu AT ucsc DOT edu
LinkedIn

Github

Publications and preprints

T. Nakamura-Zimmerer, Q. Gong, and W. Kang. Neural network optimal feedback control with enhanced closed loop stability. arXiv: 2109.07466 [math.OC], 2021.

T. Nakamura-Zimmerer, Q. Gong, and W. Kang. QRnet: optimal regulator design with LQR-augmented neural networks. IEEE Control Systems Letters, 2021.

W. Kang, Q. Gong, and T. Nakamura-Zimmerer. Algorithms of data development for deep learning and feedback design. Physica D: Nonlinear Phenomena, 2021.

T. Nakamura-Zimmerer, Q. Gong, and W. Kang. Adaptive deep learning for high-dimensional Hamilton-Jacobi-Bellman equations. SIAM Journal on Scientific Computing, 2021.

T. Nakamura-Zimmerer, Q. Gong, and W. Kang. A Causality-free neural network method for high-dimensional Hamilton-Jacobi-Bellman equations. American Control Conference, 2020.

T. Nakamura-Zimmerer, D. Venturi, Q. Gong, and W. Kang. Density propagation with characteristics-based deep learning. arXiv: 1911.09311 [math.DS], 2019.

Selected presentations

Improving reliability in neural network optimal feedback control design. SIAM Conference on Control and its Applications, Spokane WA, 2021.

QRnet: optimal regulator design with LQR-augmented neural networks. American Control Conference, New Orleans LA, 2021.

Supervised learning for optimal feedback controller design. NorCal Control Workshop, Stanford CA, 2021.

Data-driven approximation of high-dimensional flow maps and probability distributions using deep neural networks. SIAM Conference on Computational Science and Engineering, Spokane WA, 2019.

Deep learning for probability density and flow map approximation of high dimensional nonlinear systems. Bay Area Scientific Computing Day, Sandia National Laboratories, Livermore CA, 2018.

Data-driven computational optimal control for nonlinear systems under uncertainty. SIAM Annual Meeting, Portland OR, 2018.

Copyright Tenavi Nakamura-Zimmerer, 2021. Website under construction.