AM Introductory remarks
PM Panel discussion
Unsupervised Anomaly Detection and Reinforcement Learning for Navy-Relevant Scenarios
Dr. Ratna Khatri and Dr. Colin C. Olson
In this talk we will discuss two machine learning-based architectures for Navy applications. In the first part, we present a new robust learning paradigm for unsupervised anomaly detection. Our approach combines batch stochastic gradient descent algorithms and loss functions tailored for anomaly detection problems with data models comprised of either manifold learning techniques, autoencoders, or generative adversarial networks. We show that our framework demonstrates a clear improvement over state-of-the-art techniques on established benchmark data sets for unsupervised multivariate anomaly detection. In the second part, we discuss a new learning strategy using deep reinforcement learning to design better countermeasure and counter-countermeasure strategies for aerial engagement scenarios. But our method requires extensive use of a computationally-demanding, high-fidelity, physics-based aerial engagement simulation. As such, we describe our efforts to accelerate agent training via parallelization and show that within two hours of training we can generate agents that outperform state-of-the-art, hand-designed tracking algorithms.
Dr. Ratna Khatri is the Jerome and Isabella Karle Distinguished Scholar Fellow in the Optical Sciences Division at the U.S. Naval Research Laboratory in Washington D.C and an affiliate of the Center for Mathematics and Artificial Intelligence. She earned her Bachelors in Mathematics in spring 2016 and Ph.D. in Mathematics in spring 2020 from George Mason University. She has participated in numerous prestigious summer internships including SURF fellowship at National Institute of Standards and Technology, NSF Mathematical Sciences Graduate Internship Program Fellowship as well as the Givens Associate Research Fellowship at Argonne National Laboratory. She was the selected participant of the 2020 Rising Stars in Computational & Data Sciences at the Oden Institute. Her research interests are in inverse problems, deep learning and PDE-constrained optimization. Her current work focuses on algorithmic development of deep learning architectures using optimal control with applications in image processing.
Dr. Colin C. Olson earned a B.S. in Mechanical Engineering from Colorado State University in 2003 and the M.S. and Ph.D. degrees in Structural Engineering from the University of California, San Diego in 2005 and 2008, respectively, where he was a recipient of the Los Alamos National Laboratory and National Defense Science and Engineering Graduate Fellowships. He was a National Research Council Postdoctoral Fellow at the Naval Research Laboratory where he is now a staff research scientist in the Optical Sciences Division. He was awarded a DoD Laboratory University Collaboration Initiative (LUCI) grant in 2017 to support work in image processing and compressed sensing. His research interests include computational imaging, optics, reinforcement learning, nonlinear dynamics, novel solutions to inverse problems, and the application of machine learning and deep neural networks to signal processing and image analysis.
Mathematical and Computational Science Research at NIST
Dr. Günay Doğan and Dr. Justyna Zwolak
Founded in 1901, the National Institute of Standards and Technology (NIST) is a non-regulatory federal agency within the U.S. Department of Commerce. Its mission is to promote U.S. innovation and industrial competitiveness by advancing measurement science, standards, and technology in ways that enhance economic security and improve our quality of life.
The Information Technology Laboratory (ITL) is one of six major organizational units that make up the NIST Labs. Its purpose is to cultivate trust in information technology and metrology. ITL consists of seven technical divisions, including the Applied and Computational Mathematics Division (ACMD). ACMD provides leadership within NIST in the use of applied and computational mathematics to solve technical problems arising in measurement science and related applications.
ACMD researchers are involved in a wide variety of scientific projects and collaborations, ranging from physical modeling, scientific computing, machine learning, and data science. In this presentation, we will describe some of our work in ACMD. This will include the development of computational tools for image and shape analysis and the application of machine learning to enhance and control quantum experiments.
More information about the ACMD division at NIST can be found at https://www.nist.gov/itl/math and in the annual activity report , which provides and overview of all ongoing projects: https://nvlpubs.nist.gov/nistpubs/ir/2020/NIST.IR.8306.pdf
Dr. Günay Doğan is a Mathematician in the Applied and Computational Mathematics Division (ACMD) at the National Institute of Standards and Technology (NIST) in Gaithersburg, MD. He received an M.Sc. degree in Computer Science, and a Ph.D. degree in Applied Mathematics and Scientific Computing from the University of Maryland at College Park. After graduation, he did postdoctoral work on inverse problems at the University of Pennsylvania, and on computational modeling of materials at NIST. In the following years, he continued his research on image analysis and scientific computing as a Research Scientist at Theiss Research, also collaborating with NIST as a guest researcher. He has been a Staff Scientist in ACMD at NIST since 2020.
Dr. Günay Doğan’s research is broadly in the areas of image analysis, data science and scientific computing. To solve problems in these areas, he develops numerical tools for partial differential equations, variational problems, and optimization. Recent problems that he has focused on are image segmentation, the problem of identifying objects and regions and images, and shape analysis, i.e. quantitative comparison of shapes and geometries in given data.
Dr. Justyna Zwolak is a Scientist in the Applied and Computational Mathematics Division at National Institute of Standards and Technology in Gaithersburg, MD. She received an M.Sc. in Mathematics from The Faculty of Mathematics and Informatics, Nicolaus Copernicus University, and a Ph.D. in Physics from the Faculty of Physics, Astronomy and Informatics, Nicolaus Copernicus University, in Toruń, Poland. She subsequently was a research associate in the Department of Physics at Oregon State University, at the STEM Transformation Institute at Florida International University, and an assistant research scholar in the Joint Center for Quantum Information and Computer Science at University of Maryland on College Park, MD.
Her research pursuits range from quantum information theory and machine learning to complex network analysis to mathematics and physics education. In her current work, Justyna uses machine learning algorithms and artificial intelligence, especially deep convolutional neural networks, to enhance and control quantum system and quantum computing platforms. In particular, she is investigating methods to automatically identify stable configurations of electron spins in semiconductor-based quantum computing. She is also developing a complete software suite that enables modeling of quantum dot devices, training recognition networks, and -- through mathematical optimization -- auto-tuning experimental setups. Success in this endeavor will eliminate the need for heuristic calibration and help scale up quantum computing into larger quantum dot arrays.
Adventures as an Applied Mathematician at The Aerospace Corporation
Dr. Kathryn E. Brenan
My career as an industrial mathematician has been shaped by my work at The Aerospace Corporation, which operates a nonprofit Federally Funded Research and Development Center for the U.S. government. I have strived to solve real-world problems of interest to my customers in the defense and national security arenas thru the development and application of numerical algorithms.
In my talk I will describe a particular image processing algorithm I found fascinating from the point of view of the mathematics it utilizes. A basic knowledge of linear algebra, discrete fast Fourier transforms, and iterative algorithms is sufficient background for this talk. I will conclude my talk with a discussion on my work environment, and how it compares to working in academia.
I am currently a Senior Project Engineer in the Engineering & Technology Group at The Aerospace Corporation. My research & work projects have involved the development of numerical algorithms for signal & image processing applications and for differential-algebraic equations arising in trajectory optimal control problems. In the last couple years, my interests have broadened to include Monte Carlo simulations and data analysis. My work has required a solid understanding of numerical linear algebra, numerical integration, and optimization methods.
I received my PHD in Applied Mathematics & Numerical Analysis from UCLA. I am a co-author of the book, “Numerical Solution of Initial-Value Problems in Differential-Algebraic Equations.” Over the years, I have strived to stay involved in SIAM in a variety of ways, including serving on the Board of Trustees and the Council, by giving talks on my work and organizing mini-symposia for SIAM conferences, and by engaging in mentoring events.
(this abstract and bio have been approved for public release: OTR-2013-0320152229)
Efficiently connecting your research to relevant problems
Dr. Jerry TW Kim and Dr. Burhan Necioglu
This talk provides an overview of The MITRE Corporation which is a not-for-profit company that operates several federally funded research and development centers (FFRDCs). MITRE’s mission-driven teams are dedicated to solving problems for a safer world. Through our public-private partnerships and federally funded R&D centers, we work across government and in partnership with industry to tackle challenges to the safety, stability, and well-being of our nation. We tackle some of the toughest problems facing the nation, in such areas as national defense, infrastructure, cyber strategies, communications, electromagnetics, acoustics, big data, artificial intelligence, machine learning, aviation security, technology policies, etc. Consequently, MITRE performs applied research to develop new capabilities and technology solutions that are not easily attainable from industry or think tanks. Our technology transfer program seeks to mature technologies that would be best adopted by industry and government. MITRE welcomes partnership with academia to bridge many of the high technology risk areas where the shared research would not only advance the current state of the art in mathematics and the physical sciences but also bring new perspectives to current problems. Also, in this talk, we will discuss one a machine learning application in electromagnetics.
Approved for Public Release, Distribution Unlimited. Public Release Case Number 21-1703
©2021 The MITRE Corporation. ALL RIGHTS RESERVED
Dr. Jerry Kim received the B.S. degree in mathematics from the United States Naval Academy, Annapolis, MD, the M.S. degree in physics from the Naval Postgraduate School, Monterey, CA, and the M.S. and Ph.D. degrees in mathematics from the Rensselaer Polytechnic Institute, Troy, NY. He is currently the Science and Technology Outcome Leader at The MITRE Corporation for the Naval Division. He has previously worked at the Tactical Electronic Warfare Division and the Radar Division of the Naval Research Laboratory and has served in U.S. Navy as a Surface Warfare Officer (SWO) and an Engineering Duty Officer (EDO). His current interests include time reversal algorithms, singularity expansion method, electronic warfare, waveform design, radar signal processing techniques, and communications. He is an IEEE and SIAM member.
Dr. Burhan F. Necioglu is a signal processing engineer with The MITRE Corporation. He received his Ph.D. in Electrical Engineering from the Georgia Institute of Technology in 1999, his M.S. in Electrical Engineering from Boston University in 1992, and his B.S. in Electrical Engineering from the Middle East Technical University in 1989. His research interests focus on machine learning and unsupervised learning/machine discovery, and their applications to multi-dimensional and multi-modal sensor signals in diverse and novel application domains.