Efrain H Gonzalez

Postdoctoral Appointee

Postdoctoral Appointee

ehgonza@sandia.gov

Biography

Over the past five years, my research has primarily been centered around two fields: dynamic system modeling and computer vision. With regards to the former, my focus has been on the development of the theory and application of new dynamic mode decomposition (DMD) methods. With regards to the latter, my research has focused on the use of neural networks for automatic target recognition (ATR). The ATR research has allowed me to explore concepts such as explainability, interpretability, transfer learning, out-of distribution analysis, and topological data analysis.

Although the above topics have been the primary focus of my research, I have also had the opportunity to conduct research on Bayesian networks, analysis of list-mode data, and neuromorphic computing. My work involving Bayesian networks explored whether providing additional information about relationships between parents and children in the network would lead to networks that better represented the joint distribution of the variables. My research involving list-mode data explored whether incorporating time information into a classification model led to better performance on a radioisotope identification task. Regarding the neuromorphic computing research that I was a part of, the goal was to determine whether the compilation of neuromophic algorithms could be improved through the use of certain streaming graph partitioning algorithms.

All of that to say, I enjoy tackling interesting problems and learning about different fields of research.

Education

University of South Florida – Ph.D. Mathematics (concentration in Statistics) – August 2023

Florida International University – M.S. Statistics – May 2018

University of Florida – B.S. Physics & B.S. Statistics – May 2015

Publications

Published/Accepted Journal Publications

  1. L. Patel, E. H. Gonzalez, R. J. Kamm, and A. J. Hill, “Radioisotope Identification with List-Mode Gamma-Ray Data: Assessing the Value of Temporal Information Applied to Radioisotope Identification”, Data Science in Science, Published https://www.tandfonline.com/doi/full/10.1080/26941899.2025.2597582#abstract
  2. E. H. Gonzalez, B. Russo, P. Laiu, and R. Archibald, “StOKeDMD: Streaming Occupation Kernel Dynamic Mode Decomposition“, Applied Mathematics for Modern Challenges, Published https://www.aimsciences.org/article/doi/10.3934/ammc.2024022
  3. E. H. Gonzalez, M. Abudia, M. Jury, R. Kamalapurkar, and J. A. Rosenfeld, “The Kernel Perspective on Dynamic Mode Decomposition“, Transactions on Machine Learning Research, Published https://openreview.net/forum?id=sIR8xV7hGl
  4. C. Yoo, E. H. Gonzalez, Z. Gong, and D. Roy “A Better Mechanistic Understanding of Big Data through an Order Search Using Causal Bayesian Networks,” Big Data and Cognitive Computing, Published https://doi.org/10.3390/bdcc6020056
  5. S. B. Park, C. K. Chung, E. H. Gonzalez, and C. Yoo “Causal inference network of genes related with bone metastasis of breast cancer and osteoblasts using causal Bayesian networks,” Journal of Bone Metabolism, Published https://doi.org/10.11005/jbm.2018.25.4.251

Published/Accepted Conference Papers

  1. E. H. Gonzalez, C. R. Hall, C. J. Kramer, C. K. Madsen, C. M. Vineyard, and J. Adams, “Replacing Subspace Tracking Methods with Dynamic Mode Decomposition,” IEEE Aerospace Conference 2026, Accepted
  2. W. M. Severa, F. Wang, Y. Ho, F. Rothganger, A. Daram, and E. H. Gonzalez, “Benchmarking Spiking Network Partitioning Methods on Loihi 2“, Great Lakes Symposium on VLSI 2025, Published https://dl.acm.org/doi/10.1145/3716368.3735294
  3. J. Bauer, E. H. Gonzalez, W. M. Severa, C. M. Vineyard, “Benchmarking transferability metrics for SAR ATR“, SPIE Defense+Commercial Sensing, Published https://doi.org/10.1117/12.3014031
  4. J. Bauer, E. H. Gonzalez, W. M. Severa, C. M. Vineyard, “SAR ATR analysis and implications for learning“, SPIE Defense+Commercial Sensing, Published https://doi.org/10.1117/12.3014042
  5. E. H. Gonzalez, L. Avazpour, R. Kamalapurkar, and J. A. Rosenfeld, “Modeling Partially Unknown Dynamics with Continuous Time DMD,” American Control Conference, Published https://doi.org/10.23919/ACC55779.2023.10156424