Automated Infrastructure and Dependency Detection via Satellite Imagery and Dependency Profiles
Summary of LDRD efforts to be presented at LDRD poster presentation.
97 MATHEMATICS AND COMPUTING↗
SEARCH · Search NASA
Results for “Machine Learning, Object Detection, Explainability”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Automated Infrastructure & Dependency Detection via Satellite Imagery and Dependency Profiles Project Summary & Profession Background
This presentation was presented to N&HS SAC community on July 21 2022.
97 MATHEMATICS AND COMPUTING↗
Performance Comparison of Object Detection Networks for Shrapnel Identification in Ultrasound Images
Ultrasound imaging is a critical tool for triaging and diagnosing subjects but only if images can be properly interpreted. Unfortunately, in remote or military medicine situations, the expertise to interpret images can be lacking. Machine-learning image interpretation models that are explainable to the end user and deployable in real time with ultrasound equipment have the potential to solve this problem. We have previously shown how a YOLOv3 (You Only Look Once) object detection algorithm can be used for tracking shrapnel, artery, vein, and nerve fiber bundle features in a tissue phantom. However, real-time implementation of an object detection model requires optimizing model inference time. Here, we compare the performance of five different object detection deep-learning models with varying architectures and trainable parameters to determine which model is most suitable for this shrapnel-tracking ultrasound image application. We used a dataset of more than 16,000 ultrasound images from gelatin tissue phantoms containing artery, vein, nerve fiber, and shrapnel features for training and evaluating each model. Every object detection model surpassed 0.85 mean average precision except for the detection transformer model. Overall, the YOLOv7tiny model had the higher mean average precision and quickest inference time, making it the obvious model choice for this ultrasound imaging application. Other object detection models were overfitting the data as was determined by lower testing performance compared with higher training performance. In summary, the YOLOv7tiny object detection model had the best mean average precision and inference time and was selected as optimal for this application. Next steps will implement this object detection algorithm for real-time applications, an important next step in translating AI models for emergency and military medicine.
60 APPLIED LIFE SCIENCES↗
A unique, ring-like radio source with quadrilateral structure detected with machine learning
ABSTRACT We report the discovery of a unique object in the MeerKAT Galaxy Cluster Legacy Survey (MGCLS) using the machine learning anomaly detection framework astronomaly. This strange, ring-like source is 30′ from the MGCLS field centred on Abell 209, and is not readily explained by simple physical models. With an assumed host galaxy at redshift 0.55, the luminosity (1025 W Hz−1) is comparable to powerful radio galaxies. The source consists of a ring of emission 175 kpc across, quadrilateral enhanced brightness regions bearing resemblance to radio jets, two ‘ears’ separated by 368 kpc, and a diffuse envelope. All of the structures appear spectrally steep, ranging from −1.0 to −1.5. The ring has high polarization (25 per cent) except on the bright patches (<10 per cent). We compare this source to the Odd Radio Circles recently discovered in ASKAP data and discuss several possible physical models, including a termination shock from starburst activity, an end-on radio galaxy, and a supermassive black hole merger event. No simple model can easily explain the observed structure of the source. This work, as well as other recent discoveries, demonstrates the power of unsupervised machine learning in mining large data sets for scientifically interesting sources.
Astronomy & Astrophysics↗
Linking repeat lidar with Landsat products for large scale quantification of fire-induced permafrost thaw settlement in interior Alaska
The permafrost–fire–climate system has been a hotspot in research for decades under a warming climate scenario. Surface vegetation plays a dominant role in protecting permafrost from summer warmth, thus, any alteration of vegetation structure, particularly following severe wildfires, can cause dramatic top–down thaw. A challenge in understanding this is to quantify fire-induced thaw settlement at large scales (>1000 km 2 ). In this study, we explored the potential of using Landsat products for a large-scale estimation of fire-induced thaw settlement across a well-studied area representative of ice-rich lowland permafrost in interior Alaska. Six large fires have affected ~1250 km 2 of the area since 2000. We first identified the linkage of fires, burn severity, and land cover response, and then developed an object-based machine learning ensemble approach to estimate fire-induced thaw settlement by relating airborne repeat lidar data to Landsat products. The model delineated thaw settlement patterns across the six fire scars and explained ~65% of the variance in lidar-detected elevation change. Our results indicate a combined application of airborne repeat lidar and Landsat products is a valuable tool for large scale quantification of fire-induced thaw settlement.
54 ENVIRONMENTAL SCIENCES↗
A Review on the State of the Art of Machine Learning and Satellite Imaging: Detecting Scene Changes in Selected Nuclear Fuel Cycle Datasets
The timely detection of clandestine nuclear facilities is one of the greatest challenges faced by the International Atomic Energy Agency’s (IAEA). Idaho National Laboratory is currently applying machine learning (ML) to existing satellite imagery (SI) datasets to find facilities within the nuclear fuel life cycle, with primary focus placed on identifying critical predecessor (i.e., fuel fabrication and fuel enrichment) and successor (i.e., nuclear power plants) facilities. This could provide a satellite image methodology that the IAEA could leverage to discover clandestine facilities. The work presented in this paper describes the evolution of a workflow developed by this team for object detection related to critical infrastructure by expanding that workflow for the purpose of identifying nuclear fuel cycle components and automating dependency assessments. This will be done using two methods housed within a single pipeline. The first method involves implementing a DenseNet161 convolutional neural network to classify the images and explain the results using Local Interpretable Model-Agnostic Explanations (LIME). The second method implements You Only Look Once version 5 (YoloV5), to detect objects within images, provide a probability for the detection, and provide a bounding box that corresponds to the object of interest. The results of this work are anticipated to provide a clear picture of this portion of the nuclear fuel cycle and perform as a stand-alone tool for image assessment that can be expanded to additional fuel cycle components and implemented in international safeguards and national security domains. This capability addresses the IAEA’s need to detect undeclared nuclear materials and activities within a state while encompassing the entire nuclear fuel cycle.
97 MATHEMATICS AND COMPUTING↗
How Low Can You Go? Using Synthetic 3D Imagery to Drastically Reduce Real-World Training Data for Object Detection
Deep convolutional neural networks (DCNNs) currently provide state-of-the-art performance on image classification and object detection tasks, and there are many global security mission areas where such models could be extremely useful. Crucially, the success of these models is driven in large part by the widespread availability of high-quality open source data sets such as Image Net, Common Objects in Context (COCO), and KITTI, which contain millions of images with thousands of unique labels. However, global security relevant objects-of-interest can be difficult to obtain: relevant events are low frequency and high consequence; the content of relevant images is sensitive; and adversaries and proliferators seek to obscure their activities. For these cases where exemplar data is hard to come-by, even fine-tuning an existing model with available data can be effectively impossible. Recent work demonstrated that models can be trained using a combination of real-world and synthetic images generated from 3D representations; that such models can exceed the performance of models trained using real-world data alone; and that the generated images need not be perfectly realistic (Tremblay, et al., 2018). However, this approach still required hundreds to thousands of real-world images for training and fine tuning, which for sparse, global security-relevant datasets can be an unrealistic hurdle. In this research, we validate the performance and behavior of DCNN models as we drive the number of real-world images used for training object detection tasks down to a minimal set. We perform multiple experiments to identify the best approach to train DCNNs from an extremely small set of real-world images. In doing so, we: Develop state-of-the-art, parameterized 3D models based on real-world images and sample from their parameters to increase the variance in synthetic image training data; Use machine learning explainability techniques to highlight and correct through targeted training the biases that result from training using completely synthetic images; and Validate our results by comparing the performance of the models trained on synthetic data to one another, and to a control model created by fine-tuning an existing ImageNet-trained model with a limited number (hundreds) of real-world images.
97 MATHEMATICS AND COMPUTING↗
Report on Next-Gen AI for Proliferation Detection Workshop: Domain-Aware Methods
The emergence of artificial intelligence (AI) and machine learning (ML) in the modern world has impacted nearly every application imaginable. This includes nuclear proliferation detection, which offers the potential to improve existing capabilities as well as create new ones. Proliferation detection seeks to detect and characterize attempts by state and non-state actors to acquire nuclear weapons or associated technology, materials, or knowledge. Such a mission is vitally important for global stability and security but is notoriously difficult. By leveraging advances in AI, exciting opportunities exist to enhance the proliferation detection regime. The Data Science and AI portfolio within the National Nuclear Security Administration’s Office of Defense Nuclear Nonproliferation Research and Development (DNN R&D) seeks to leverage the capabilities of the Department of Energy’s (DOE’s) national laboratories and other partners to develop AI systems that can accomplish otherwise impossible tasks in support of proliferation detection. As part of its efforts, the portfolio has created a series of workshops on Next-Gen AI for Proliferation Detection to help define the requirements for suitable AI systems, share successful research and best practices, and foster connection and understanding between the relevant parties including researchers and end-users. Each workshop in the series focuses on a specific and critical aspect of AI to enable it to accomplish proliferation detection objectives. The first workshop focused on explainability techniques; the second workshop and the topic of this report, covers methods for incorporating domain awareness into AI. The Next-Gen AI for Proliferation Detection Workshop: Domain-Aware Methods took place virtually over two days in February 2021 and included four keynote presentations, 22 technical presentations, and a concluding panel. The presentations, discussions, and workshop findings are summarized in this report.
97 MATHEMATICS AND COMPUTING↗
SEARCHING FOR MESONIC DARK MATTER WITH THE HEAVY PHOTON SEARCH EXPERIMENT
Several highly-sensitive astrophysical experiments over the past couple of decades have demonstrated that the current abundance of visible Standard Model matter cannot explain galactic rotation curves, the expansion history of the Universe, or the apparent warping of light in empty space. Instead, one finds strong agreement with this body of experimental results upon positing the existence of an invisible particulate field, dark matter. Namely, a cold, weakly interacting dark matter component can explain all these phenomena. A number of accelerator-based experiments have been developed to search for the weak couplings/interactions of these particles, many of them concentrating on particle models with masses of tens to thousands of GeV. A relatively new, well-motivated model is a dark sector coupled to the Standard Model via a dark photon. The current abundance of dark matter can be obtained if one assumes that dark matter is coupled to light by a MeV to GeV particle with a U(1) symmetry. The parameter space of these models remains largely unexplored because they are difficult to probe experimentally. In this thesis, I analyze data from the Heavy Photon Search (HPS) detector, whose two detector halves closely surround the electron beam, providing acceptance to far-forward boosted interactions. This forward acceptance to highly boosted particles yields unprecedented sensitivity to MeV-scale invariant masses. I exhaustively optimize the offline reconstruction of the HPS detector. Each reconstruction object, from Silicon Vertex Tracker hits to tracks, is studied to maximize acceptance of dark matter events. I then use the 2021 run data to search for one model of dark-photon-mediated matter, the Strongly Interacting Massive Particle (SIMP). SIMP models provide self-interacting dark matter candidates that can form bound states resembling dark mesons. HPS can detect SIMPs through the decay of a dark vector boson (either a dark ¿ or ¿) into e+e- pairs. I obtain exclusion contours for SIMPs using both an optimized cuts-based selection and a machine-learning-based selection, advancing our knowledge of the nature of dark matter.
O'Dwyer, Rory [Stanford Univ., CA (United States).↗
Big Data Synchrophasor Monitoring and Analytics for Resiliency Tracking (BDSMART)
This report contains key findings from a project titled Big Data Synchrophasor Monitoring and Analytics for Resiliency Tracking (BDSMART), which was carried out through a collaborative effort of a team of researchers from Texas A&M Engineering Experiment Station, Temple University, and Quanta Technology, LLC. The in-kind support came from OSIsoft (acquired by AVEVA), which provided their PI Historian software to demonstrate the use case of streaming PMU data. The first section of the report describes the project goals and objectives related to the development of Machine Learning (ML) models capable of detecting and classifying events by processing phasor measurements captured in the field by Phasor Measurement Units (PMUs). The data for this study was contributed by the utilities/ISOs from the Western and Eastern interconnects and ERCOT, further referred to as Interconnect B (IC B), Interconnect A (IC A), and Interconnect C (IC C), respectively. The approach that the BDSMART Research Team proposed and the key research tasks defined by the team are outlined in this section. The next section describes the technical approach. We first discuss the data constraints related to the PMU measurements and data interpretation constraints imposed by the data contributors. They provided neither the topological information of the grid nor PMU placement locations and captured recorded data at very few locations in the system with the reporting rate of either 30 or 60 fps. The recordings are mostly positive sequence voltage, frequency, and ROCOF, and in some limited cases, three-phase voltages and currents. We then reflect on the bad data issues that stem from poor recording practices and vague definitions of the PMU status bits to supposedly be used for bad data identification. Finally, the data discovery points to imprecise time stamps with incomplete event start/end time, as well as inconsistent and incomplete event labeling, which combined make the implementation of the data models using supervising learning quite challenging. Following the data discovery study, we hypothesize that because the IC B data has the most complete labels, we should focus our model development on that data and then test it on data from other interconnects. We also define the common metrics used to evaluate the results from the ML algorithm tests. We concluded this section by summarizing the common ML models we used and explaining how we implemented and tested them. The issues from this section are expanded in the Training Dataset Report from this project. The final section of this report deals with the accomplishments and conclusions. As the accomplishments, we formulate the problem we are solving and what is achieved by solving the problem. We then reflect on each of the analytics tools we developed and point out the performance of each tool when applied to solving the mentioned problems. We reference this work for further details to the papers we published on each tool. In the conclusions, we give recommendations on how to improve future PMU recording practices to facilitate the ML algorithm implementation and guidance for the future standardization work aimed at clarifying the ambiguities associated with the PMU status bits. We finally list future tasks that can bring about further improvements in the proposed algorithms. The issues from this section are expanded in the Training, and Test Dataset Report filed at the project completion date.
97 MATHEMATICS AND COMPUTING↗