Five early career researchers were recently selected for the Institute for Digital Research and Education (IDRE) Postdoctoral Fellowship Awards. Applicants were nominated by their faculty sponsors and the recipients were then selected by a committee comprised of several IDRE Executive Committee members. They were hand-picked based on their proposals’ merit featuring innovative and computationally based solutions to research problems with a potentially broad impact to the campus.
These IDRE fellowships will provide partial support for research conducted in 2019-20 and related to the IDRE programs. The IDRE fellows will have direct access to and interact with technologists within the IDRE Research Technology Group and relevant IDRE affiliated faculty and researchers.
The UCLA postdoctoral fellows includes Dr. J. Harry Caufield, Integrative Life Sciences; Dr. Matthew Jacobs, Mathematics; Dr. Masaki Nakada, Computer Science; Dr. Zhi Jackie Yao, Electrical and Computer Engineering; and Dr. Khalid Youssef, Biomedical Engineering.
John Caufield’s research focuses on how to best learn about physiologically meaningful phenomena described within unstructured data. He is specifically interested in developing and applying text mining methods to biomedical documents relevant to rare disease presentations, with an emphasis on the vocabulary and background knowledge particular to prominent fields such as cardiovascular disease research. He integrates data mining and computational approaches that have yet to be widely adopted in biomedicine, including natural language processing, knowledge graphs, and graph-based inference. “The overall goal of my work is to develop a platform for the analysis of protein interactions and modifications likely to occur in the course of rare disease presentations. Such a platform will provide support for deep phenotyping of specific disease etiologies and contribute evidence to new diagnostic approaches.” John Caufield states, “The support and opportunities provided through the IDRE Fellowship are invaluable to pursuing this integrative strategy.”
Provided By: John Caufield
Matthew Jacobs’ broad area of research is the calculus of variations, which studies minimization problems over spaces of functions. This is an extremely useful tool in physics, as many physical processes are driven by energy minimization. At the moment, Matthew is very interested in free boundary problems. These kinds of problems show up whenever there is an interaction between materials with different physical properties, such as oil mixing with water. In his research, he will try to show that solutions to these types of problems exist, and attempt to find efficient numerical methods to actually compute the solutions. IDRE’s fellowship will support, “the funding to help [him] travel to the international congress on industrial and applied mahematics in Spain.”
Provided By: Matthew Jacobs
Masaki Nakada’s research comprises of a computer simulation model and associated software system for biomimetic human sensorimotor control. Sensorimotor control is an essential aspect of making virtual humans autonomous. It advances the state-of-the-art in computer graphics animation technology for use by the motion picture, game, virtual/augmented/mixed reality, and education industries, particularly in applications requiring realistic, self-animating virtual humans. It applies machine learning techniques from Artificial Intelligence, specifically deep learning, to create a “brain” for a biomechanically simulated human musculoskeletal model that is actuated by numerous contractile muscles. By synthesizing its own training data, the virtual human automatically learns efficient, online, active visuomotor control of its eyes, head, and limbs in order to perform nontrivial tasks involving the foveation and visual pursuit of target objects, coupled with visually-guided limb-reaching actions to intercept the moving targets, as well as to carry out drawing and writing tasks.
Provided by: Masaki Nakada
Zhi Jackie Yao’s Magnetic materials could enable the design of new classes of radio frequency-based components. These components could then be used in everything from smartphones to tiny, implantable health-monitoring devices. Nevertheless, the lack of powerful design tools dealing with interactions between magnetic materials and EM waves hinders the in-depth understanding of the role of magnetism in dynamic and high-frequency systems. The tools also have limits in the design of consumer electronics. Yao’s research goal is to design a new tool to model how magnetic materials, important in sending and receiving in communications devices, interact with incoming radio signals, down to nanometer scales. This computational tool addresses the existing modeling problems by giving electronics designers a clear path toward figuring out how potential materials would be best used in communication devices. The computational tool is based on a method that jointly solves well-known Maxwell’s equations, which describe how electricity and magnetism work and the Landau–Lifshitz–Gilbert equation, which describes how magnetization moves inside a solid object. Zhi Jackie Yao says that, “With the resources that IDRE provides, I am much better equipped to investigate the underexplored territory.”
Provided By: Zhi Jackie Yao
Khalid Youssef’s research targets the computational bottlenecks in the mathematical operations that previously prevented parallelization of 2nd order training for large scale artificial neural networks. He has developed methods that allow exponential reduction in computational time and memory resources required for performing 2nd order optimization. He is currently working on optimizing the implementation of his methods for a number of different computational environments, including memory optimized computer servers with high number of CPU cores, GPU servers with multiple GPUs, and multiple separate machines working in parallel, in order to expand the scale of applications that can benefit from 2nd order training. He is collaborating with several labs to develop custom machine learning systems for their specific applications. Khalid Youssef says that, “The IDRE fellowship is an opportunity to start new collaborations and connect with other researchers with overlapping interests.”
Provided By: Khalid Youssef
MORE ABOUT THE FELLOWS
Harry Caufield is a postdoctoral fellow in the NIH HeartBD2K Center of Excellence at UCLA, where he works with Prof. Peipei Ping of UCLA’s departments of Physiology, Medicine, and Bioinformatics. Before joining UCLA, Dr. Caufield earned his PhD in Integrative Life Sciences at Virginia Commonwealth University, where he studied microbial protein interactions and developed intuitive methods for working with large protein interaction networks. He continues to have an active interest in learning about biological relationships hidden within disparate data sources, particularly those with a direct impact on human health and disease.
Matthew Jacobs was born and raised in Brooklyn, NY. He received a B.A. in math at Columbia University in 2011. Then received a Ph.D. in math from the University of Michigan in 2017. His thesis advisor was Selim Esedoglu. For the last two years, he has been a postdoc in the UCLA math department and his postdoc mentors are Stan Osher and Andrea Bertozzi.
Masaki Nasada obtained his Bachelor’s degree in applied physics in 2009 and his Master’s degree in pure and applied physics in 2011 from Waseda University in Tokyo, Japan. He then obtained his Ph.D. in computer science in 2017 from University of California Los Angeles, and his current research is centered on artificial life to create a virtual human with biomimetic human vision and a neuromuscular control system. His research has led to the development of a biomimetic human system as a bottom up approach, which gives us the ability to realize the complete human system in software and robots.
Zhi (Jackie) Yao is currently a postdoctoral researcher in the Electrical and Computer Engineering (ECE) Department at University of California, Los Angeles (UCLA). She obtained the M.S. degree in 2014 and the Ph.D. degree in 2017, both in the ECE Department at UCLA. Her main work during the graduate and postdoc researches is the development of a numerical solver unifying dynamic electromagnetics (EM), nonlinear magnetic spin dynamics, and acoustics. This solver is under progress of licensing and enabling public access, and is one of the key contributions to the sustainability of the NSF ERC center for Translational Applications of Nanoscale Multiferroic Systems (TANMS) headquartered at UCLA. Her work has been funded by sources such as NSF and DARPA.
She has received multiple academic honors, such as the 1st place Best Student Paper in International Microwave Symposium, IEEE Antennas and Propagation Society Doctoral Research Grant, Qualcomm Innovation Fellowship, Outstanding Master’s Thesis at ECE UCLA and 1st place in PhD exam at ECE UCLA.
Khalid Youssef is a postdoctoral researcher with a Ph.D. in Biomedical Engineering. His research involves developing new machine learning algorithms, and their application in a variety of interdisciplinary fields such as image & signal processing, robotics, control, fluid dynamics, and medical imaging. Some applications he developed include a new generation of bioreactors that utilize artificial neural networks to dynamically control mechanical forces in cell growth environment, state of the art medical image denoising capable of learning noise statistics from images, and data driven radio frequency transmitter identification. Khalid’s work has pushed the state of the art in 2nd order ANN training algorithms, and his ongoing research targets the implementation of novel 2nd order-based ANN training methods to large-scale and big-data applications.