MD-SOAR Maryland Shared Open Access Repository
Not a member yet
34521 research outputs found
Sort by
A Catalog of Filaments in the Central Molecular Zone
No full textWe present a catalog of 1.28 GHz radio filaments observed by MeerKAT over the innermost 200 pc of the Galaxy (roughly ±1.5°), which includes the central molecular zone (CMZ). The catalog is generated by repurposing software developed for the automated detection of filaments in solar coronal loops. There are two parts to the catalog. The first part, the main catalog, provides a point-by-point listing of locations and basic observational properties along each detected filament. The second part is a summary catalog which provides a listing of mean, median, or total values of various properties for each filament. Tabulated quantities include position, length, curvature, brightness, and spectral index. The catalogs contain a heterogeneous mix of filamentary structures, including nonthermal radio filaments (NRFs), and parts of supernova remnants (SNRs) and thermally emitting regions (e.g. H II regions). We discuss criteria for selecting useful subsamples of filaments from the catalogs, and some of the details encountered in examining filaments or selections of filaments from the catalogs.Work by R.G.A. was supported by NASA under award numbers 80GSFC21M0002 and 80GSFC24M0006. This work is also partially supported by the grants AST 2305857 from the NSF. We thank the anonymous referee for their help in improving the manuscript.http://arxiv.org/abs/2510.1249
MCPS Plans High School Enrichment Overhaul, But Can it Deliver?
No full textMCPS is reorganizing its high school specialty programs, ending countywide recruitment for its marquee magnets, and introducing 30 separate programs and 84 pathways across 6 regions. Sunil Dasgupta talks with parent volunteers Rebekah Kuschmider and Audra Dove, members of the Opportunity Design Team organized by MCPS to engage the community, about whether MCPS can really deliver on its promises on equity and learning. Plan details: https://www.montgomeryschoolsmd.org/curriculum/academic-programs-analysis/. Music by Drew Pictures and the Lead Extras.https://open.spotify.com/episode/4j65l4tUkmXxAJXf8SZ1l
Comparing machine learning approaches for predicting the success of ICSI treatment: A study on clinical applications
No full textIntracytoplasmic Sperm Injection (ICSI) is widely used to treat almost all forms of male infertility and to overcome fertilization failure. While ICSI is a powerful procedure, it's also considered quite expensive, which means couples and clinicians have to make informed decisions about whether or not to proceed with this treatment. About 10,036 patient records, 46 attribute sets, and one label column that indicates the success or failure of pregnancy after the ICSI treatment were used to conduct this research. The data were gathered from Razan infertility center in Palestine. The ICSI dataset contains only clinical features that are known prior to deciding on ICSI treatment. The dataset contains 46 features, 5 of the independent features have categorical values, 12 are numerical, 3 are string, and 26 are binary. Based on the results, RF algorithm achieved the highest AUC score of 0.97, followed by the NN with a score of 0.95, and the RIMARC algorithm with a score of 0.92. AUC is a widely used metric for evaluating the performance of binary classification models. Therefore, judging by the AUC scores, it appears that RF algorithm outperformed the other two algorithms in terms of the evaluated metric. The method employed in our analysis demonstrates considerable promise, practicality, and generalizability, driving advancements in fertility treatments and ultimately improving the chances of couples achieving their desired family goals.https://www.sciencedirect.com/science/article/pii/S266652122500007
Editorial: Statistical and Nonlinear Physics Crosses a Threshold
No full textWe are delighted to announce the transition of the American Physical Society's (APS) Topical Group on Statistical and Nonlinear Physics (GSNP) to a Division (DSNP [1]) as a result of a steady growth in membership. Divisions are the largest units representing subfields within APS, with increased opportunities for scientific sessions at meetings, broader participation, and better representation within APS governance. We find it very fitting that Physical Review E, the APS journal with which the scope of our activities best aligns, has graciously invited us to mark this event in its pages. PRE and GSNP were born in the same period—the journal in 1993 and the topical group only a few years later in 1997—with a similar unifying intent: the interfaces of the field were rapidly multiplying, and there was a need to anchor many disparate systems to their strong disciplinary roots within physics. There is also a strong overlap with the scope of topics that PRE is home to [2]: Statistical physics; nonlinear dynamics and chaos; networks and complex systems; biological physics; soft matter including polymers, liquid crystals, and granular materials; mechanics, interfaces, and films; fluid dynamics; plasma physics; computational physics, machine learning, and artificial intelligence. The shift in emphases within GSNP have also mirrored the shift in focus of PRE. When PRE started in 1993, its subtitle was “Statistical physics, plasmas, fluids, and related interdisciplinary topics.” In 2000, the subtitle of PRE added soft matter physics. Thereafter, new areas of research, such as network science and machine learning have become a significant part of PRE. GSNP has also expanded its focus to accommodate emerging topics in these years.https://journals.aps.org/pre/abstract/10.1103/PhysRevE.112.02000
Dr. Arielle Kuperberg's Public Comment on the U.S. Department of Education's Proposed Title IV Regulations
No full textHello, my name is Arielle Kuperberg PhD, I'm an Associate Professor of Sociology at the University of Maryland, Baltimore County, and I’ve been doing research on student loans for around ten years. This research has included analysis of national datasets along with surveys and interviews of college graduates with and without loans who my collaborators and I have followed over time for up to eight years past graduation.https://scholars.org/contribution/dr-arielle-kuperbergs-public-comment-u
Urinary Biomarkers and Joint Cognition-Gait Trajectories: Findings from the Health, Aging, and Body Composition (ABC) Study: SA-PO1113
No full textBackground:
Emerging urinary biomarkers can unlock new links between kidney tubular injury and cognitive-physical function.
Methods:
Using the Health ABC Study, we examined baseline urinary biomarkers—uromodulin, alpha-1 microglobulin (α1M), amino-terminal propeptide of type-III procollagen (PIIINP), neutrophil gelatinase-associated lipocalin (NGAL), interlukin-18 (IL-18), and kidney injury molecule-1 (KIM-1)—and joint cognition-gait trajectories among baseline high-function older adults. IL-18 and KIM-1 were measured in n=1902 participants; uromodulin, α1M, PIIINP and NGAL were measured in a random subcohort (n=502). Group-based trajectory analysis of 20m usual gait speed and modified Mini-Mental State up to year 10 revealed three groups: high cognitive-physical function (Group 1), stable cognition/declining gait (Group 2), and rapid joint decline (Group 3).
Results:
After adjusting for covariates in separate models (Model-1A), higher α1M (p=0.043) and KIM-1 (p=0.005) concentrations were related to worse trajectories (Table). Further adjustment for estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (UACR) and c-reactive protein (CRP) (Model-1B) attenuated estimates; KIM-1 was not quite significant (p=0.059). In a fully adjusted model of all urinary markers (Model-2), only KIM-1 was significant (p=0.012), but the set of urinary biomarkers was jointly significant (p=0.022).
Conclusion:
KIM-1 was robustly related to cognitive-gait decline, and may link kidney tubular injury with aging-related functional outcomes.Other NIH Support - National Institute of Aging, National Institute of Neurological Disorders and Strokehttps://journals.lww.com/jasn/fulltext/2025/10001/urinary_biomarkers_and_joint_cognition_gait.3764.asp
Sub-wavelength optical lattice in 2D materials
No full textRecently, light-matter interaction has been vastly expanded as a control tool for inducing and enhancing many emergent nonequilibrium phenomena. However, conventional schemes for exploring such light-induced phenomena rely on uniform and diffraction-limited free-space optics, which limits the spatial resolution and the efficiency of light-matter interaction. Here, we overcome these challenges using metasurface plasmon polaritons (MPPs) to form a sub-wavelength optical lattice. Specifically, we report a “nonlocal” pump-probe scheme where MPPs are excited to induce a spatially modulated AC Stark shift for excitons in a monolayer of MoSe₂, several microns away from the illumination spot. We identify nearly two orders of magnitude reduction for the required modulation power compared to the free-space optical illumination counterpart. Moreover, we demonstrate a broadening of the excitons’ linewidth as a robust signature of MPP-induced periodic sub-diffraction modulation. Our results will allow exploring power-efficient light-induced lattice phenomena below the diffraction limit in active chip-compatible MPP architectures.S., M.J.M., d.G.S.-F., c.J.F., l.X.,and M.h. were supported by ARO W911nF2510066, dARPA hR00112530313, and hR00112490310.l.G. and Y.Z. were supported by ARO W911nF2510066, the national Science Foundation (nSF)dMR-2145712, and the department of energy (dOe) de-Sc-0022885 grants. this research usedQuantum Material Press (QPress) of the center for Functional nanomaterials (cFn), which is a USdepartment of energy, Office of Science User Facility, at Brookhaven national laboratory undercontract no. de-Sc0012704. K.W. and t.t. acknowledge support from the JSPS KAKenhi (grant nos.20h00354, 21h05233, and 23h02052) and World Premier international Research center initiative(WPi), MeXt, Japan, for hBn synthesishttps://www.science.org/doi/full/10.1126/sciadv.adv202
The Impact of Pandemic Era Health Care Policy on Uninsured Non-Elderly Adult Americans in an ACA Medicaid Expansion and a Non-Expansion State
No full textD.P.A. -- The University of Baltimore, 2025Dissertation submitted to the School of Public Affairs of The University of Baltimore in partial fulfillment of the requirements for the degree of Doctor of Public AdministrationThis paper reviews pre-pandemic, pandemic era, and post-pandemic health care policy and health care access barriers for uninsured non-elderly adult Americans between two cities. One in a Medicaid expansion state and one in a non-expansion state. A comparison will be made of institutional health care access policy, local, state, and federal policy, access disparities, and social determinant factors among uninsured non-elderly adult Americans. Factors will be identified using quantitative analysis of U.S. Department of Health and Human Services (DHHS). A summary of factors influencing the uninsured in the presence of pandemic era local, state, and federal policy. A quasi-experimental mixed methods analysis will be provided to show the impact of pre-, intra-, and post-COVID-19 pandemic era health policy on the uninsured living in Baltimore City and the City of Milwaukee.
Original literature shows a significant increase in access to health care for the uninsured following the implementation of Patient Protection and Affordable Care Act (ACA) in 2010 with gradual decreases in access overtime resulting in a negative impact on health care delivery for uninsured non-elderly Americans living in the two subject cities. Local-, state-, and federal level health care policy demonstrate resources available to offset costs of caring for the uninsured with comparative examples of two cities. community organizations, leveraging state, federal and private funding, facilitate health care for uninsured Americans experiencing homelessness. Comparative data will be examined to show if there is a significant disparity in access to health care between the states of Maryland and Wisconsin and specifically focusing on Baltimore City and the City of Milwaukee
Achieving Fairness for Free in Artificial Intelligence Systems via Bayesian Optimization
No full textAs artificial intelligence (AI) has become increasingly embedded in high-stakes decision-making systems, from criminal justice to hiring and healthcare, ensuring fairness is critical. However, prevailing assumptions suggest that improving fairness inevitably compromises predictive performance. This thesis challenges that assumption by exploring whether fairness can be achieved “for free”, that is, without sacrificing accuracy, through the use of Bayesian Optimization (BO). We frame fairness-aware machine learning as a black-box constrained optimization problem, where the goal is to maximize fairness while satisfying a predefined accuracy threshold. We incorporate fairness metrics such as the p%-rule and use BO to efficiently search over hyperparameters, treating both the model architecture and fairness-accuracy trade-off parameter (λ) as part of the search space. The core of our method involves using Gaussian Processes as a surrogate model, along with acquisition functions like Expected Improvement (EI) and Upper Confidence Bound (UCB) to balance exploration and exploitation. Extensive experiments on benchmark datasets (e.g., COMPAS, Adult Census, and Bank) demonstrate that BO outperforms traditional grid search in both runtime efficiency and fairness-performance trade-off. In many cases, our BO-based method finds models that improve fairness without any measurable drop in accuracy, achieving the “fairness for free” phenomenon. This work contributes a practical and scalable approach to tuning fairness-sensitive models in black-box settings and lays the groundwork for further deployment of fair AI in real-world applications. By reframing fairness optimization as a sample-efficient search problem, we help bridge the gap between ethical AI principles and technical feasibility
Rivaling Transformers: Multi-Scale Structured State-Space Mixtures for Agentic 6G O-RAN
No full textIn sixth-generation (6G) Open Radio Access Networks (O-RAN), proactive control is preferable. A key open challenge is delivering control-grade predictions within Near-Real-Time (Near-RT) latency and computational constraints under multi-timescale dynamics. We therefore cast RAN Intelligent Controller (RIC) analytics as an agentic perceive-predict xApp that turns noisy, multivariate RAN telemetry into short-horizon per-User Equipment (UE) key performance indicator (KPI) forecasts to drive anticipatory control. In this regard, Transformers are powerful for sequence learning and time-series forecasting, but they are memory-intensive, which limits Near-RT RIC use. Therefore, we need models that maintain accuracy while reducing latency and data movement. To this end, we propose a lightweight Multi-Scale Structured State-Space Mixtures (MS³M)¹ forecaster that mixes HiPPO-LegS kernels to capture multi-timescale radio dynamics. We develop stable discrete state-space models (SSMs) via bilinear (Tustin) discretization and apply their causal impulse responses as per-feature depthwise convolutions. Squeeze-and-Excitation gating dynamically reweights KPI channels as conditions change, and a compact gated channel-mixing layer models cross-feature nonlinearities without Transformer-level cost. The model is KPI-agnostic -- Reference Signal Received Power (RSRP) serves as a canonical use case -- and is trained on sliding windows to predict the immediate next step. Empirical evaluations conducted using our bespoke O-RAN testbed KPI time-series dataset (59,441 windows across 13 KPIs). Crucially for O-RAN constraints, MS³M achieves a 0.057 s per-inference latency with ∼0.70M parameters, yielding 3-10x lower latency than the Transformer baselines evaluated on the same hardware, while maintaining competitive accuracy.http://arxiv.org/abs/2510.0525