26800 research outputs found
Sort by
INTERACTIONS AND APPLICATIONS OF OPTICAL ANGULAR MOMENTUM IN MAGNETO-OPTICAL MATERIAL
No full textThis dissertation explores the interactions and applications of optical angular momentum within magneto-optical materials. Beginning with a theoretical and experimental analysis of multiple reflection and refraction phenomena within magneto-optical material. We derive and verify the dependence of refractive indices on optical spin angular momentum and magneto-optical magnetization, resulting in nonreciprocal elliptical and linear polarization beam splitting effects and wavevector discretization. We fabricate a magnetless slab waveguide isolator with minimal optical loss. Extending our study to optical orbital angular momentum, we introduce a perturbation to the electronic transition model in bismuth-substituted iron garnets. We demonstrate that this perturbation leads to nonreciprocal modification of the orbital angular momentum state-specific refractive index and reduction of Faraday rotation. This is experimentally verified using a Mach-Zehnder interferometer. As a culmination of the results, we leverage the orbital angular momentum state-specific Faraday rotation reduction to create a magneto-optical-based orbital angular momentum state selection algorithm. This orbital angular momentum state selection and modulation algorithm has the same scaling as Grover\u27s algorithm, presenting a scalable, high-accuracy probabilistic computing architecture
A Special Class of Pure O-Sequences
No full textThe pure O-sequences of the form (1, a, a, …) are classified
A Joint Geometric Topological Analysis Network (JGTA-Net) for Detecting and Segmenting Intracranial Aneurysms
Get PDFObjective: The rupture of intracranial aneurysms leads to subarachnoid hemorrhage. Detecting intracranial aneurysms before rupture and stratifying their risk is critical in guiding preventive measures. Point-based aneurysm segmentation provides a plausible pathway for automatic aneurysm detection. However, challenges in existing segmentation methods motivate the proposed work. Methods: We propose a dual-branch network model (JGTANet) for accurately detecting aneurysms. JGTA-Net employs a hierarchical geometric feature learning framework to extract local contextual geometric information from the point cloud representing intracranial vessels. Building on this, we integrated a topological analysis module that leverages persistent homology to capture complex structural details of 3D objects, filtering out short-lived noise to enhance the overall topological invariance of the aneurysms. Moreover, we refined the segmentation output by quantitatively computing multi-scale topological features and introducing a topological loss function to preserve the correct topological relationships better. Finally, we designed a feature fusion module that integrates information extracted from different modalities and receptive fields, enabling effective multi-source information fusion. Results: Experiments conducted on the IntrA dataset demonstrated the superiority of the proposed network model, yielding state-of-the-art segmentation results (e.g., Dice and IOU are approximately 0.95 and 0.90, respectively). Our IntrA results were confirmed by testing on two independent datasets: One with comparable lengths to the IntrA dataset and the other with longer and more complex vessels. Conclusions: The proposed JGTA-Net model outperformed other recently published methods (\u3e 10% in DSC and IOU), showing our model\u27s strong generalization capabilities. Significance: The proposed work can be integrated into a large deep-learning-based system for assessing brain aneurysms in the clinical workflow
Technical feasibility, technoeconomic, and life cycle assessment of CO2 sequestration using domestic nickel mine tailings: A direct ex-situ hydrothermal approach
No full textDirect ex-situ mineral carbonation offers a pathway for permanent CO2 sequestration through reactions with silicate minerals. While conventional mining and processing of feedstocks are energy-intensive and costly, the utilization of pre-ground mine tailings rich in CO2-reactive silicates may provide a more economical alternative. This study evaluates the technical feasibility, technoeconomic assessment (TEA), and life cycle assessment (LCA) of direct aqueous mineral carbonation using domestic nickel mine tailings. Experimental results demonstrated that 16–35 kg CO2 per ton of mine tailings could be sequestered at 125–185 °C without additional comminution. Further particle size reduction, or the use of the slime fraction, increased CO2 uptake to 46–91 kg CO2 per ton. Mineralogical characterization confirmed the carbonation of olivine and a fraction of pyroxene to form siderite and magnesite, while plagioclase and quartz remained largely inert. The TEA/LCA analysis revealed a sequestering cost of \u3e 100 per ton CO2 and GWP \u3c1) would require ultrafine tailings with ≥80 % olivine content and the use of low-carbon heat source. This study represents the first integrated evaluation of both the technical feasibility and technoeconomic scenarios of CO2 sequestration using direct ex-situ carbonation route for nickel mine tailings
Exceptional Points and Lasing Thresholds: When Lower-Q Modes Win
No full textOne of the most fundamental questions in laser physics is the following: Which mode of an optical cavity will reach the lasing threshold first when gain is applied Intuitively, the answer appears straightforward: When a particular mode is both temporally well confined (i.e., exhibits the highest quality factor) and experiences initially the largest increase of the modal gain, it is naturally expected to lase first. However, in this Letter, we demonstrate that this intuition can fail in surprising ways. Specifically, we show that in the presence of non-Hermitian degeneracies, known as exceptional points, the expected mode hierarchy can be dramatically altered. These spectral singularities can give rise to counterintuitive mode switching, where a mode with a lower quality factor and initially smaller increase of modal gain reaches the lasing threshold ahead of a more favorable competitor. Remarkably, this effect can occur even under spatially uniform pumping, underscoring the subtle and profound influence of non-Hermitian physics on lasing dynamics
SIAMESE: Stealing Fine-Tuned Visual Foundation Models via Diversified Prompting
No full textVisual foundation models, characterized by their robust generalization and adaptability, serve as the basis for a wide array of downstream tasks. When fine-tuned for specific tasks, these models encapsulate confidential and valuable task-specific knowledge, making them prime targets for model stealing (MS) attacks. While recent efforts have exposed MS threats in practical scenarios such as data-free and hard-label contexts, these attacks predominantly target traditional victim models trained from scratch. Fine-tuned visual foundation models, pre-trained on vast and diverse datasets and then fine-tuned on downstream tasks, present significant challenges for traditional MS attacks to extract task-specific knowledge. In this paper, we introduce an innovative MS attack, named SIAMESE, to steal fine-tuned visual foundation models under black-box, data-free, and hard-label settings. The core approach of SIAMESE involves constructing a stolen model using a foundation model that is efficiently and concurrently fine-tuned with multiple diversified soft prompts. To integrate the knowledge derived from these prompts, we propose a novel and tractable loss function that analyzes the output distributions while enforcing orthogonality among the prompts to minimize interference. Additionally, a unique alignment module enhances SIAMESE by synchronizing interpretations between the victim and stolen models. Extensive experiments validate that SIAMESE outperforms state-of-the-art baseline attacks over 10% in accuracy, exposing the heightened vulnerability of fine-tuned visual foundation models to MS threats
Uncertainty-Aware Pavement Roughness Forecasting Using Adaptive Conformal Prediction
No full textAccurate forecasting of future pavement conditions is essential for effective asset management and enables data-driven maintenance planning. While recent studies have demonstrated the potential of Artificial Intelligence (AI)-based methods to predict future pavement roughness, their application remains limited due to challenges in quantifying uncertainty. Most machine learning models rely on point-based accuracy metrics, which fail to capture performance variability across the full range of pavement conditions—particularly underrepresented or extreme scenarios where prediction errors tend to increase. To address this limitation, this study applies an adaptive conformal prediction—a forecasting framework that integrates seamlessly with AI algorithms to generate prediction intervals with targeted confidence levels. Specifically, we enhance the conventional conformal framework by developing a quantile-based adaptive strategy that dynamically adjusts interval widths according to the difficulty of each prediction. The adaptive Jackknife + approach consistently outperforms other methods, delivering 90.3% empirical coverage and an efficiency score of 0.588 at the 90% confidence level, while reducing the average interval width by more than 20% compared to its regular counterpart. Conditional-coverage analysis across pavement, traffic, and climatic features further demonstrates that the adaptive framework maintains stable and well-calibrated uncertainty across diverse operating conditions
CDP-KDNet: Curriculum-Guided Dynamic Pruning and Knowledge Distillation for Resource-Efficient Ultrasound Elastography
No full textIn recent years, convolutional neural network (CNN)-based optical flow models for motion estimation have been applied to radio-frequency (RF) ultrasound and B-mode (BM) data, demonstrating excellent performance. However, their architectures result in intricate network structures with a large number of parameters, posing challenges for deployment on resource-constrained devices. This paper proposes a novel approach that integrates dynamic pruning, knowledge distillation, and curriculum learning for model compression. The proposed method substantially reduces the complexity of deep learning models (i.e., memory demands and computational costs) while minimizing performance degradation. The teacher network was initially developed based on the Unsupervised Motion Estimation CNN (UMEN-Net). Subsequently, we developed a sub-network to reduce the number of parameters, referred to as DP-Net, and applied the proposed training techniques to obtain the final model, CDP-KDNet. The CDP-KDNet model was evaluated on simulated, phantom, and in vivo ultrasound data. Compared to DP-Net and other lightweight CNNs, CDP-KDNet achieves superior Signal-to-Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR) for axial strain estimation across all tested datasets. Its performance closely matches that of the teacher network while utilizing only 45.3% of the parameters and 67.8% of the floating-point operations. Additionally, as an unsupervised model, CDP-KDNet does not require ground-truth labels during training, rendering it a promising approach for ultrasound motion estimation
DIGITAL LIGHT PROCESSING OF POLYMER-DERIVED CERAMIC PRECURSORS FOR THE FABRICATION OF SILICON OXYCARBIDE COMPOSITES
No full textCeramic composites are highly desirable for aerospace and extreme environment applications due to their exceptional thermal stability, mechanical strength, and oxidation resistance. Polymer-derived ceramics (PDCs) offer a versatile route to fabricating such composites, with the added advantages of tunable properties—such as electrical and thermal conductivity—and processing at lower temperatures and pressures compared to conventional sintering methods. Their polymeric origin also enables additive manufacturing via vat-based techniques. In this work, SiOC ceramic composites were synthesized from photo-curable polysiloxane precursors and shaped into complex geometries using Digital Light Processing (DLP). Photocurable groups were incorporated into the polymer backbone, enabling UV crosslinking at 405 nm. To tailor composite properties, Silicon Carbide, Hexagonal Boron Nitride, and Alumina were introduced as fillers. The effects of these additives on both printability and final microstructure were evaluated. Pyrolysis at 1000 °C converted the crosslinked structures into ceramics. The resulting composites are expected to exhibit reduced shrinkage, enhanced mechanical properties, thermal stability, and thermal conductivity, making them strong candidates for aerospace applications
Assessing Climate Vulnerabilities in Rural and Low-Income Regions: A Case Study of the Western Upper Peninsula
Get PDFLow-income and rural regions may be more vulnerable to the impacts of climate change because they face challenges in accessing resources and systemic exclusion from climate planning coupled with other geographical, geological, and social factors. Climate vulnerability assessments (CVAs) offer a logical framework for assessing the potential impacts of climate change on socio-ecological systems and provide a scientific basis for developing adaptation strategies. Using secondary data from scholarly papers, online publications, government documents, official reports, and relevant news items, the study investigates Michigan’s Western Upper Peninsula’s (WUP) vulnerability to climate change and climate variability. The study is guided by a conceptual framework highlighting the climatic and non-climatic factors influencing and impacting exposure, sensitivity, and adaptive capacity in assessing vulnerability toward resilience building and adaptation planning. The study found that access to financial resources, diversified livelihoods, and climate-literate social networks will reduce vulnerabilities in the WUP. In contrast, aging population, aging infrastructure, poverty, financial resource constraints, energy service outages, institutional mistrust, low health service density, and pollution remnants from mining legacies contribute to climate vulnerabilities in the WUP. The vulnerability of the rural and low-income community should be addressed through just and equitable policy and strategies