37200 research outputs found
Sort by
Matrix-Free High-Performance Saddle-Point Solvers for High-Order Problems in H (div)
Get PDFThis work describes the development of matrix-free GPU-accelerated solvers for high-order finite element problems in . The solvers are applicable to grad-div and Darcy problems in saddle-point formulation, and have applications in radiation diffusion and porous media flow problems, among others. Using the interpolation–histopolation basis (cf. [W. Pazner, T. Kolev, and C. R. Dohrmann, SIAM J. Sci. Comput., 45 (2023), pp. A675–A702]), efficient matrix-free preconditioners can be constructed for the -block and Schur complement of the block system. With these approximations, block-preconditioned MINRES converges in a number of iterations that is independent of the mesh size and polynomial degree. The approximate Schur complement takes the form of an M-matrix graph Laplacian and therefore can be well-preconditioned by highly scalable algebraic multigrid methods. High-performance GPU-accelerated algorithms for all components of the solution algorithm are developed, discussed, and benchmarked. Numerical results are presented on a number of challenging test cases, including the “crooked pipe” grad-div problem, the SPE10 reservoir modeling benchmark problem, and a nonlinear radiation diffusion test case
Building Partnerships Through Third-Party Facilitation: Best Practices from the Community Collaborative Initiative
Get PDFSmall minority group organisations rarely collaborate with other nonprofits owing to a lack of expertise and resources. Third-party facilitators can help these groups navigate the process of collaboration. However, the literature has largely ignored their role in the process. We address this gap by studying the challenges third-party facilitators face in the collaboration process and best practices they can apply using the Community Collaboration Initiative (CCI), a unique third-party-facilitated collaboration process working with Muslim American nonprofits
Strategies to Mitigate Electrostatic Charging During Coffee Grinding
Get PDFSummary: Coffee grinding generates electrostatically charged particles, causing clumping, spark discharge, and beyond. When brewing, the particle aggregates affect liquid-solid surface accessibility, leading to variable extraction quality. Here, we study four charge mitigation strategies. De-electrification is readily achieved by adding small amounts of water to whole beans or by bombarding the grounds with ions produced from a high-voltage ionizer. While these techniques helped reduce visible mess, only water inclusion was found to impact coffee extracts prepared as espresso. Wetting whole beans with less than 0.05 mL/g resulted in a marked shift in particle size distribution, by preventing clump formation and preventing fine particles from sticking to the grinder. This particle size shift results in at least a 15% higher coffee concentration for espresso extracts prepared from darker roasts. These findings encourage the widespread implementation of water use to de-electrify coffee during grinding with the benefit of increased coffee extraction efficiency
W86 - A Care Cascade Analysis of Opioid Use Disorder Services Provided in Jail
No full textAn Analysis of Opioid Use Disorder Services Provided in Oregon Jail
Portland State University Economic and Social Impact Study 2024
Get PDFPortland State University (PSU) serves as a vital economic and social engine for the Portland metropolitan area and the state of Oregon. With more than 20,000 students enrolled across eight distinctive colleges and schools, offering more than 200 degree and certificate programs, and employing a total workforce exceeding 3,500, PSU contributes significantly to economic, social and community vibrancy through its core educational, research and engagement activities. This study examines PSU’s economic and social contributions, educational outcomes, research activities and community engagement initiatives through both quantitative and qualitative analyses to provide a multifaceted view of the university’s impact.
Geographically, this study concentrates on the Portland Metropolitan Statistical Area (MSA)1 where PSU is located, and where the majority of its employees and students reside. Additionally, the study considers the regional contributions of PSU to the State of Oregon as one of the three largest public universities within the state. The analysis timeframe centers around Fiscal Year 2023 (July 1, 2022 to June 30, 2023) and Academic Year 2022-2023 (Fall 2022 to Summer 2023).
Following a literature and case study review, this study develops the PSU Economic and Social Impact Analysis Framework (Figure 1), expanding on the Associate of Public Land-grant Universities’ (APLU) Talent, Innovation, and Place framework (APLU, 2023) to include PSU-specific data inputs and analytical components. We then present a profile of the university within the Oregon context. PSU’s economic and social impacts are characterized and assessed through an economic contribution analysis (input-output analysis), place-based analysis and a hybrid social and community impact analysis of PSU’s educational and research activities
Book Review of, Calling Family: Digital Technologies and the Making of Transnational Care Collectives
Get PDFBook Review: Calling Family: Digital Technologies and the Making of Transnational Care Collective
DCAI: The 4th International Workshop on Data-Centric AI
Get PDFMachine learning traditionally emphasizes developing models for given datasets, but real-world data is often messy, making model improvement insufficient for enhancing performance. Data-Centric AI (DCAI) is an emerging field that systematically improves datasets, leading to significant practical ML advancements. While experienced data scientists have manually refined datasets through trial-and-error and intuition, DCAI approaches data enhancement as a systematic engineering discipline. DCAI represents a shift from focusing on models to the underlying data used for training and evaluation. Despite the dominance of common model architectures and predictable scaling rules, building and using datasets remain labor-intensive and costly, lacking infrastructure and best practices. The DCAI movement aims to develop efficient, high-productivity open data engineering tools for modern ML systems. This workshop seeks to foster an interdisciplinary DCAI community to address practical data challenges, including data collection, generation, labeling, preprocessing, augmentation, quality evaluation, debt, and governance. By defining and shaping the DCAI movement, this workshop aims to influence the future of AI and ML, inviting interested parties to contribute through paper submissions
Families at the Center: Leading for Home Visiting Systems Change
Get PDFThe Center for Coordinating Oregon Home Visiting Systems (CCOHVS) launched in February 2024 at the Center for Improvement of Child and Family Services at Portland State University. The CCOHVS functions as a program-neutral backbone organization for a coordinated system of prenatal and early childhood home visiting services. Its work will move the state closer to the vision described in the Raise Up Oregon 2.0 (RUO) strategic plan1 for early childhood systems, namely to create “equitable, integrated, accessible, inclusive, anti-racist and family-centered” early learning services, with a focus on ensuring this system meets the needs of pregnant people and families with infants and toddlers.
To reach this goal, the CCOHVS team uses an inclusive, relationship-focused approach to build on emerging innovations that are building and changing existing systems. Working with local, regional, and state agency partners the CCOHVS team provides needed capacity and support to achieve home visiting system goals, including ensuring family leadership in home visiting system transformation at the state and local levels.
To this end, the CCOHVS team prepared this Learning Brief, which summarizes key principles and practices that Oregon’s Home Visiting System Initiatives (HVSI) governance and advisory groups should consider in their approaches to engaging families in leadership and decision-making to inform and shape home visiting systems change
Strengthening Capacity in Industrial Revolution 4.0 (IR 4.0): A Case Study on Training Product-Service Systems Between the US and Vietnam
Get PDFThis case study analyzes how value is created in the development process of a collaborative training product-service system (PSS) designed for provincial government officials in Vietnam. This initiative, part of the USAID Strengthening Provincial Capacity (USAID SPC) project, encapsulates a series of training solutions designed to empower officials with key skills in engineering and technology management, including the knowledge to improve global value chains, upgrade technology infrastructure and sustainable production models. A primary driver for the “four-day course delivery”, the PSS analyzed in this study, was to bolster economic competitiveness at the provincial level in the context of Industrial Revolution 4.0 (IR 4.0). This case study offers lessons on stakeholder-value mapping and recommendations for future approaches to training courses development that enhances capacity development. These trainings aligned with the demands of balancing organizational development and public value creation in the context of developing countries transitioning to a knowledge economy, such as Vietnam. The learnings from this case study may provide a blueprint for similar endeavors in other countries navigating the complexities of digital transformation of public services
Techniques for Modeling Ocean Soundscapes: Detailed Description for Wind Contributionsa
Get PDFWind over the ocean creates breaking waves that generate air-filled bubbles, which radiate underwater sound. This wind-generated sound is a significant component of the ocean soundscape, and models are essential for understanding and predicting its impact. Models for predicting sound pressure level (SPL) from wind have been studied for many years. However, the terminology and definitions behind modeling approaches have not been unified, and ambiguity has led to differences in predicted SPL. The 2022 Ambient Sound Modeling Workshop was organized to compare ambient sound modeling approaches from different researchers. The main goal of the workshop was to quantify differences in predicted SPL and related quantities for different approaches and, to the extent possible, determine the cause of the differences for a specific, well-defined scenario. Results revealed a variation of approximately 6 dB across different research groups, with differences reaching up to 10 dB in some cases compared to the benchmark results described in this paper. These variations stemmed from differing methodologies and underlying assumptions. In this paper, step-by-step guidance is given for modeling SPL due to wind. The workshop test case will be described, and results from the modeling approaches described here will be compared with those from the workshop participants