12466 research outputs found
Sort by
Overcoming Barriers in Front-End Web Development
Full text linkFront-end web developers frequently encounter challenges such as asynchronous behaviors and complex API interactions, often due to barriers in programming and debugging interfaces that obscure the relationship between code and its execution. These challenges, highlighted in Stack Overflow posts, emphasize the difficulties in modifying existing code and navigating through code idioms—distinct code patterns with similar semantics. For example, the semantics of System.out.println() and console.log() pertain to externalizing internal data. Our analysis of these posts has led to the categorization of API interactions into a set of idioms and the identification of associated barriers, revealing a clear need for tools that can trace, scrub, and search across the execution to locate visual elements or runtime values without losing their linkage to the corresponding code. SeeCode.Run, an innovative online IDE, directly tackles these needs with features like per-expression always-on execution recording and interactive visualizations that bring to light execution facts about code. This integration of live programming and debugging techniques provides developers with in-editor widgets for immediate execution insights, facilitating code fact-checking and continuous editing. Its omniscient search capability and scope navigation enable developers to effortlessly understand and manage the myriad of framework interactions, addressing barriers like silent invalid references, unclear scopes, and the tracking of state changes within asynchronous flows. The effectiveness of SeeCode.Run's fact-driven approach was empirically tested against CodePen, a standard in online live programming IDEs. This comparison demonstrated that developers using SeeCode.Run experienced a significant reduction in task completion time and an increase in success rates, validating the platform's potential to overcome front-end development challenges and enhance productivity through a more intuitive workflow
Hardware Implementation of the FAEST PQC Digital Signature Scheme
Full text linkQuantum computers are becoming increasingly powerful every day, and with each new
milestone, we are closer to a quantum computer capable of solving the mathematical problems
that form the basis of classical asymmetric cryptography. To prepare for this, experts
are developing new cryptographic schemes known as Post-Quantum Cryptography (PQC),
which can be implemented with classical computers and are expected to be secure against
classical and quantum computers.
This thesis provides an overview of the current state of PQC, which is in a transition
period. The first standards for key encapsulation mechanisms (KEMs) and digital signatures
have been selected but are still being finalized, while an additional call to find different
candidates for PQC digital signatures has recently been published.
This situation presents a challenge for the cryptographic community because efforts must
be divided between finalizing decisions on existing standards and starting the evaluation of
new candidates.
The contributions of this thesis include a detailed examination of FAEST, a promising
symmetric-based candidate from the additional round of PQC digital signatures. This thesis
also provides valuable new data on FAEST’s performance when implemented in hardware,
allowing for a more comprehensive analysis and comparison with other candidates
BENCHMARKING AND CREATION OF A DEEP LEARNING COMPUTER VISION DATASET IN THE TAXONOMIC REVISION OF THE PALO SANTO TREE HERBARIUM SPECIMENS
Full text linkPremise of the studyHerbarium collections have historically been used as a manual data source for taxonomies and there are large collections that have been digitized since the start of the 21st century. Few have been curated for computer vision problems beyond image classification, and even fewer are available to the public. In this study, I investigated the steps and effort it takes both human and computer to create a publically-available data set that can be used for image classification, object detection, and segmentation. Secondly, while benchmarking the data set, I investigated the relationship between the algorithm class, data augmentation, and network size. Our hypothesis for the first research question is that, despite the tools and technology out there to help expedite the process, that it still requires a substantial amount of human effort to complete the task. Secondly, I believe that there will be signals within the data, and that there is a relationship between the class of algorithm, data augmentation, and network size. Methods: A team of three people, one botanist, a data scientist, and research assistant contributed to the creation of this benchmark data set. Data was collected both from historical and new field samples and digitized in Washington D.C. B. graveolens and B. penicillata were selected as the benchmarking species, as they are rare tree species to reduce variability and ensure and were able to collect most of the known herbarium samples around the world. Supervisely, a web-based tool to scale image annotations, was used to annotate objects such as plant features, and to track efforts [1]. Python 3.7 was used for post- processing and creation of the data sets from Supervisely [2]. Descriptive statistics for univariate and bivariate were conducted using mean (sd), frequency counts, and appropriate statistical graphing. Deep learning models were created via PyTorch and torchvision for classification (B.graveolens and B. penicillata), semantic segmentation (plant pixel versus non-plant pixel), and instance segmentation (index card, stamp, measurement bar, color bar, barcode, compound leaf, terminal leaf, and woody material), and evaluated over different augmentation strategies and network sizes [3], [4]. Results: 1,081 images were taken of 794 biologically unique samples. 115,279 actions across 1,200 human hours were taken to annotate data. 15 types of objects were annotated across the images, resulting in 19,834 objects. For classification, a model without pretrained weights based on VGG-11 performed best with an accuracy of .727, AUC of 0.685, specificity .742, and precision .265, as compared to the second best model of ResNet- 18 .565, .665, .530, and .201. FCN-100, without upsampling and a dilation factor of 1, recorded the highest dice coefficient 0.7566. MASK R-CNN V1.0 performed better for both the non-biologic and biological models with MAP full .846 and .736, respectively optimized at 50x and 25x upsampled. Classification and semantic segmentation of larger networks for ResNet and MASK R-CNN in comparison performed worse regardless of the upsample size.Conclusion: Although a substantial amount of time was spent creating this data set, it is a modular process, and the toolkits of the 21st century have made this a process tractable with a small research team. The amount of training data is different for each algorithm class despite it being a single set of images. The smaller the training data, the greater the impact of upsampling, and conversely network size. I hope that this newly-created benchmarking data and findings help researchers interested in computer vision in herbarium research make progress toward bridging known gaps in the field
Project 4b: Exploring Variations in Gut Microbiome Networks among Patients with Chronic Kidney Disease (CKD)
Full text linkThis project included one student researcher who wishes to remain anonymous.Objective: The objective of this project is to analyze the gut microbiome ecosystems of patients with isolated CKD, CKD with comorbid liver diseases, or CKD with comorbid diabetes using a network-based approach, and to compare the differences in gut microbiome networks among these three groups of patients
A Simulation Study of SiC/Diamond & β-Ga2O3/Diamond Heterojunction Diodes
Full text linkWide and ultra-wide Band-Gap semiconductors allow for advanced electronics progress due to featuring physical and electrical properties beyond that of Silicon, and are essential for advancements in design and production of electric vehicles (EVs) and power electronics. Diamond has unique properties that are well-suited for high power applications. Combining the cost effectiveness of various semiconductors, Silicon Carbide and Beta-Phase Gallium Oxide, we can improve upon homogeneous designs using Diamond for a PN Heterojunction. Sentaurus TCAD was used to design 4H-SiC/Diamond and β-Ga2O3/Diamond Heterojunction Diodes with various dopant concentrations in the range of 1e16 cm-3 to 1e18 cm-3, along with defects, and with temperatures ranging from 270K to 400K to test diode performance, primarily in I-V characteristics, and with Breakdown voltage under reverse bias testing currently in progress. The dimensions, dopant choices, dopant concentrations, defects, and temperature ranges applied were chosen based on an extensive literature review to better reflect real-world parameters. Overall, the devices showed expectedly wide bandgaps with favorable properties, such as in the electric field, electrostatic potential, current density figures. They displayed saturation region MOSFET I-V characteristics beyond roughly 30V voltages in forward bias; this behavior was unexpected, but may serve the applications of power electronics well given all aspects of the device’s performance, especially given the prospect of future maturity of Diamond fabrication for electronics, paired with the benefits of a heterogeneous design
The Impact of Induced Fatigue on Driver Vigilance
Full text linkAs automated vehicles (AVs) become more prevalent on the roadway, drivers are becoming interested in the benefits that they may provide. An AVs potential to perform driving tasks for tired individuals is appealing for those who are at risk of drowsy driving. This manuscript covers three studies on this topic. In Study One, participants were placed in 30-minute simulated drives after performing fatiguing tasks. Survey and physiological data were analyzed to examine the impact of acute fatiguing tasks. Study Two used the same experimental tasks with participants who experienced high levels of fatigue in their day-to-day life. Study Three sought to gather the opinions of these fatigued groups regarding the use of AVs to mitigate drowsy driving. Collectively, these studies aimed to inform decision-makers of the factors that should be considered in a driver’s choice to use automation while fatigued
“Nobody Can Take Away Your Knowledge” STEM Identity Testimonios of First-Generation-In-College Latinx Undergraduate Students
Full text linkThis narrative study examines the testimonios of first-generation-in-college Latinx STEM undergraduate students using LatCrit as a theoretical lens. Four Latinx STEM undergraduate students who identify as first-generation-in-college were interviewed about how their STEM identity initially developed, what challenges they negotiated between their other identities and their STEM identity, and the experiences and identities that strengthened their STEM identity. Three unstructured interviews or pláticas were conducted with each participant. Seven common themes were identified: (a) participants have a passion and interest in their respective STEM fields, (b) childhood, middle, and high school experiences and opportunities are important for STEM identity development, (c) challenges with being a Latina in STEM, (d) challenges with doubting their STEM competence, (e) being recognized by others strengthens their STEM identity, (f) having support from family and peers strengthens their STEM identity, and (g) sense of STEM competence strengthens their STEM identity. Recommendations discussed in the conclusion of this study ask K-12 educators and administrators to ensure that all students receive equitable access to advanced-level mathematics and science courses and to provide outside-of-school STEM experiences that are accessible to all students. Undergraduate institutions also have a responsibility to ensure the success of Latinx STEM students
Ground-based Light Curve Follow-up Validation observations of TESS object of interest TOI 5907.01
No full text“This study analyzes the candidate exoplanets and their host stars, the data of which was gathered by the NASA Transiting Exoplanet Survey Satellite, and the ground-based observations from the George Mason University Satellite. In particular, we focused on the exoplanet TOI 5907.01, which orbits the star TOI 5907. We used this data to plate-solve and perform multi-aperture photometry to analyze light curves. The goal of this study is to confirm the planetary nature of this exoplanet based on the transit in the light curve. We were unable to confirm, which suggests that TOI 5907.01 was a false positive and not a real exoplanet.
Vultural Differences: A Multispecies Ethnography of Black Vulture Predation on Cattle in Virginia
Full text linkThis dissertation addressed the growing conflict between cattle farmers and black vultures in the state of Virginia using a multispecies ethnography framework. Using a mixed method design, I built a better understanding of the phenomenon of black vulture predation on livestock including its frequency, risk factors, and how perception of this issue is affected by human dimensions such as values, attitudes, and cultural factors. Specific methods included the VLDQ-R (Vulture Livestock Damage Questionnaire Revised) survey, key informant interviews, content analysis of relevant media, and GIS (Geographic Information Systems) mapping and statistical modeling of predation events, potential attractants, and vulture populations. This resulted in a detailed understanding of this understudied human-wildlife conflict that can be used to guide policy and practice. It further developed the multispecies ethnography framework by demonstrating how techniques and theory from multiple disciplines can be combined to understand and include both the human and nonhuman agents in a conflict or other interaction
Routing Optimization on LEO Satellite Networks Using Orchestrator
Full text linkIn today’s telecommunications, wireless systems based on 5G technology have emerged as dominant due to their unparalleled speed and low latency. In the last few years, we have witnessed a process of integration of ground-based telecommunication systems with satellite communication networks. This integration of terrestrial and satellite systems uses Low-Earth Orbit (LEO) satellites, which offer the lowest latency among orbital configurations due to their proximity to Earth. Several constellations of LEO satellites have already been launched, and more are being planned. In a combined terrestrial-satellite system, the signal between two end users may travel through a chain of LEO satellites placed on different orbits. Management of such communications involves finding an optimal route, thus leading to the optimal path-finding issues.The research described in this dissertation addresses the issue of the optimized routing algorithms and the properties of systems using the optimal routes. We assume there is centralized communication control at the ground-based terminals. The function of the centralized controller is performed by an orchestrator utilizing real-time satellite data from Two-Line Element (TLE) files and ephemeris data. This makes it possible to construct the LEO network layout for a given orbital period and calculate optimal paths between satellite nodes. To address the issue of path-finding in a constellation of satellites, we performed a study involving the application of three established path-finding algorithms—Bellman-Ford, Dijkstra, and A-star—focusing on their computational complexity and time complexity in the context of satellite constellations. We assumed a model involving N satellites placed on M orbital planes, with Q satellites that are evenly distributed on each orbital plane. In the specific simulation examples we assumed N = 54, M = 6, Q = 9 or alternatively N= 108, M=12, Q=9. Our results indicate that the A-Star algorithm significantly outperforms both Dijkstra and Bellman-Ford algorithms in terms of efficiency, providing superior performance for dynamic LEO satellite networks in minimizing latency and improving throughput. After establishing the A-star algorithm as the preferred path-finding process on LEO satellite networks, our study focuses on improving real-time traffic management and congestion control in dynamic LEO constellations through a central orchestrator. We enhance the A-Star algorithm within the orchestrator, where we prioritize various service-level requirements, such as latency optimization and throughput optimization, based on application-specific needs. We demonstrate how the algorithm can be enhanced to adapt to evolving satellite topologies by recalculating optimal paths in response to congestion while satisfying the service-level agreements at the same time. The results highlight the superior adaptability of the enhanced A-Star algorithm compared to traditional routing methods, making it ideal for service-oriented network orchestration in LEO satellite networks and ensuring efficient data transmission. We further continue to explore the development of effective load-balancing strategies on LEO satellite networks, specifically targeting the congestion management of Optical Inter-Satellite Links (OISLs). The study focuses on the use of the A-Star algorithm as the basis for a new load-balancing algorithm within the orchestrator that ensures efficient routing of data packets, considering Service Level Agreements (SLAs) for maximum availability and minimal latency. By integrating the microservices with the orchestrator to provide real-time updates, we enhance the algorithm to address challenges such as network topology changes, user demand fluctuations, and weather conditions affecting satellite communication. The research demonstrates how dynamically rerouting traffic and preemptively managing congestion on LEO satellite networks improves overall network performance and reduces packet loss, especially during high-demand scenarios or under adverse weather conditions. We continue to improve the efficiency of route-finding in dynamic LEO networks through the addition of caching capability within the orchestrator. By caching frequently used routes, the system reduces the need for repeated calculations, improving performance during data transmission. The study evaluates several caching strategies, including selective elimination, cache reuse after orbital periods, and predictive caching, all of which are shown to significantly reduce processing time and improve overall network efficiency. The caching strategy has been simulated not only on the constellation of 54 or 108 satellites, but also on constellations of 3060 satellites where N = 3060, M = 340, Q = 9 and 6120 satellites where N = 6120, M = 680, Q = 9. The results demonstrate the potential of caching to enhance 5G-Advanced services by lowering latency and optimizing resource usage, particularly in large-scale LEO constellations. Overall, this research explores methods to optimize time for path-finding and routing traffic over LEO satellite networks, meeting the stringent requirements of 5G while enhancing network performance and reliability