Courtesy Photo | During his internship at Air Force Research Laboratory (AFRL), Sensors Directorate at...... read more read more
Seldom does a graduate student’s research extend its influence across the Department of Defense (DoD), yet for one Science, Mathematics, and Research for Transformation (SMART) scholar, it did just that.
During his internship at Air Force Research Laboratory (AFRL), Sensors Directorate at Wright-Patterson Air Force Base, Nathan Jones began working with his SMART mentor, Rachel Kinard, Ph.D., a research mathematician at AFRL, to reproduce and refactor topological data analysis (TDA) code. TDA is used to understand and analyze data by looking at the shape and structure of a surface. It has far reaching applications such as sensor data fusion, network analysis, and predictive modeling.
For their efforts, Jones and Kinard earned the SMART Scholar and Mentor of the Year Award. This prestigious acknowledgement celebrates scholar-mentor pairs who exhibit exceptional dedication and accomplishments throughout their participation in the SMART program. As a Phase 3 scholar in the SMART Program, Kinard, now a DoD civilian, is well-prepared to apply her experience as a SMART scholar to mentor and support Jones during his internship and beyond.
Kinard discovered Jones’ prior TDA research while seeking solutions to technical challenges. Upon discovering that Jones was a SMART scholar interning at AFRL, Kinard collaborated with him and others to establish the TDA working group (TDAWG). Significantly, their collaboration transformed research in TDA into working code accessible for all DoD members.
Jones’ contributions to the TDAWG began during his summer internship. Working closely with Kinard, Jones reproduced a final technical report on TDA conducted by Paul Schrader, Ph.D. and fellow TDAWG member. Schrader’s report used TDA methods with electro-optical, infrared, and acoustic modalities in target recognition using the Air Force’s Experiments, Scenarios, Concept of Operations, and Prototype Engineering (ESCAPE) II data set. During this process, Kinard empowered Jones to apply his TDA expertise to enhance the TDA code, leading to a more efficient and widely accessible product across DoD High Performance Computing (HPC) systems.
Jones’ enhancements to Shrader’s code resulted in a two to three times increase in processing speed and improved flexibility, facilitating easier reproduction in other experiments. Jones also suggested a different method for some calculations, which made the process four times faster and allowed for larger images to be analyzed.
Jones didn’t stop there. He created a singularity recipe file, which ensures that the TDA code can be easily deployed and run on multiple DoD HPC systems, making it more accessible for other researchers within the DoD. He then shared all these improvements through a GitLab repository where others in the Department could access and use the tools.
Additionally, Jones wrote code to compute metrics for entire sets of images, making it easier to analyze large amounts of data. Finally, he ensured the code could work with other software, proving that the TDA pipeline he developed could be adapted for various data used by the DoD.
Jones attributes his successful summer to Kinard, acknowledging her role in fostering his development, delivering tangible value to the Air Force, and providing invaluable career guidance.
Heading into the fall, the mentor pair worked alongside government and academia partners on three projects to further prepare Jones for his DoD civilian career. The first project, Meshes with Algebraic Topology: TDA Evaluation and Refinement, focused on studying 3D mesh characterization and evaluation, tackling the challenge of determining whether a 3D computer-aided design model is “watertight,” a concept lacking a precise definition. Jones used TDA methods to provide a definition inspired by algebraic topology, ensuring that a mesh meets the criteria of being “watertight”.
The second project, Topological and Algebraic Data Analysis, builds upon Schrader’s previous work by incorporating additional data from the ESCAPE II dataset, including radio identification information, into the TDA pipeline. They’re also enhancing the pipeline to study how different layers of data, such as sensor data from various sources and samples of different materials, can be combined using artificial intelligence and machine learning techniques.
Finally, the third project, Sheaves for Sensor Exploitation, explores sheaves, an object from abstract mathematics, to analyze data fusion. The advantage of this approach is that it can work with various types of data and provide insights into effective methods for combining different datasets.
As a mentoring pair, Jones and Kinard have made significant contributions to the Air Force and the greater DoD. Jones’ code is available on HPC systems for all members of the DoD. Jones and Schrader presented their work to the greater AFRL community at Wright-Patterson Air Force Base. Their success has resulted in a three-year effort led by Schrader to continue improving TDA methods. The work with meshes is being submitted to the International Society for Optics and Photonics and was shared at the Naval Applications of Machine Learning Conference in March, expanding the reach of their work to the Naval Research Labs in San Diego.
Jones graduated from the University of Oklahoma in December 2023, with a SMART-sponsored master’s degree in mathematics and has begun Phase 2 of the SMART Program, his DoD civilian employment commitment, where he will continue his work with AFRL, alongside his mentor.
| Date Taken: | 04.04.2024 |
| Date Posted: | 04.04.2024 16:13 |
| Story ID: | 467831 |
| Location: | ALEXANDRIA, VA, US |
| Web Views: | 39 |
| Downloads: | 0 |
This work, SMART Alum and Scholar Awarded for Transforming Topological Data Analysis Across the DoD, must comply with the restrictions shown on https://www.dvidshub.net/about/copyright.
No keywords found.
LEAVE A COMMENT