33760 research outputs found
Sort by
BAYESIAN NETWORK AND RAILWAY TRACK DETERIORATION MODELING
No full textThe safety and reliability of the railroad infrastructure are a priority for the railroad industry.The Federal Railroad Administration requires that inspections be carried out at least twice a
week. This is particularly a challenge for the local railroad that relies on traditional inspection
methods due to the high capital cost of the track inspection machine. The large amount of data
generated by this machine is also a challenge for the local railroads to process since they lack
the required expertise. This research proposes the use of a Bayesian Network to model the
relationship between the components of the superstructure, substructure, and the geometry of
the railroad.
The proposed approach incorporates expert knowledge and a historical railway inspection
dataset to construct a Bayesian Network model that captures the causal relationships among
failure elements. A junction tree was constructed from the Bayesian Network model for exact
inference. The study further evaluates the impact of individual component deterioration on the
overall system performance.
This approach is valuable to all classes of railroads, as it helps transform the large amount of
railway data into a scalable, interpretable format for informed decision-making and risk minimization.
Results from our model demonstrate the ability of the Bayesian Network to serve as a sensitivity
analysis tool and as a mechanism to plan for scheduled maintenance operations on the railroad
A 3D GIS Web Application for Map Library Exploration
No full textSlides used to present the McKeldin Library GIS and Data Service Center's 3D Tour project at the 2025 LRIPF. For more information, contact [email protected]
Precedented: Open Digital Research Practices in Latin America
No full textThis is an Accepted Manuscript of a book chapter published by Routledge in Digital Libraries Across Continents on June 13, 2025.The broadly successful transnational scientific repositories of Latin America throw into high relief the shortcomings of “American-style” publishing and digital library models that continue to revere individual and organizational laurels over reputation and reach, and splashy features over durability. The perpetuation of the prestige economy ensures that journal subscriptions and article processing fees that cost thousands of dollars are somehow not seen as predatory in the Western world, compared to an open access publication in a trusted global repository. These priorities are still fundamentally rooted in neoliberal and imperialist thinking about scientific knowledge as a commodity produced by and for individuals, and so long as these attitudes and the policy apparatus that sustains them persist, true open access will never be attained in North America. This chapter will highlight the early history of open infrastructure and digital repositories in Latin America, their focus on regionally-produced scientific research rather than institutional or disciplinary efforts, the current state of these projects, and apply a critical lens to Western discourse about these projects and their impact. Using SciELO as a foundational example, the authors will summarize how the financial and operational sustainability of these platforms differs from the individualized and disconnected repository strategy common in the United States, which results in much stronger production and engagement with open science practices among Latin American scientists compared to their global north peers. By drawing from and reflecting upon the authors’ respective backgrounds as a Latin American bibliographer and digital librarian in elite U.S. research universities, trained in U.S. library science programs, this chapter proposes embracing an explicitly collectivist and anti-colonial approach to digital repositories. The authors will thoughtfully examine the social and political changes necessary to ignite a revolution within the heart of the prestige economy, and truly unshackle scientific knowledge from gatekeeping for-profit scholarly publishers.https://www.routledge.com/Digital-Libraries-Across-Continents/Yang-Salaz/p/book/978103264608
Language and Policy: Preservation of Minority Languages in China
No full textThe research was featured in Maryland Global on November 20, 2024: https://marylandglobal.umd.edu/about/news/fire-researchers-champion-minority-language-preservationThis paper explores the effectiveness of language preservation initiatives in China, a nation rich in linguistic diversity and home to 128 spoken languages, 107 of which are minority languages or facing endangerment. Language loss threatens cultural diversity, educational quality, social mobility, and social identity. Prior research discusses the importance of language preservation, the advantages of language policy, as well as how language policy has developed over time. Building upon that research, this research examines ineffective language policy, considering what factors may contribute to its decreased impact. This research was conducted using content analysis performed on existing literature, as well as narrative analysis on two interviews conducted with members of the affected groups. It concludes that language policy is ineffective due to the failure to account for the social and economic circumstances of the affected groups, reinforcing the necessity of inclusive language policy and preservation efforts to support cultural inclusion, identity, mobility, and education
Rediscovering Brooklyn: A Forgotton Architectural Landscape of Baltimore
No full textThis project consisted of a reconnaissance-level architectural survey of Brooklyn, a working-class Baltimore neighborhood whose significance has traditionally been overlooked by historians and preservationists. Established in the 1850s as an independent town in Anne Arundel County, and annexed by Baltimore in 1918, Brooklyn has been largely omitted from the historiography of both jurisdictions. Architecturally, Brooklyn’s “inconsistent streetscapes” have been cited as a key factor in the neighborhood’s ineligibility for listing in the National Register of Historic Places. This report aims to reframe “inconsistent streetscapes” as architectural variety that does not invalidate Brooklyn’s architectural significance but provides an opportunity to examine temporal changes to regional vernacular housing trends.
Brooklyn consists of approximately 900 acres and over 3,000 buildings. Of the extant buildings, 97% were constructed prior to 1975 and are considered historic. The neighborhood is mixed-use, with housing as the dominant building type. Through the reconnaissance survey, 3,007 dwellings were documented. The architectural survey was supplemented with archival research to better understand the historic context of the neighborhood. Five periods of development were identified based on local or national historic events that seem to have impacted Brooklyn’s built environment. The impacts include changes in development trends, construction methods, and building styles.
Each period of development illustrates Brooklyn’s evolving identity, marked by a persistent tension between urban and suburban development. While Brooklyn’s homes are modest in size and stylistic embellishments, they reflect the distillation of local and national trends through the lens of a working-class community. The “inconsistent streetscapes” are a feature of Brooklyn’s landscape that allows for further exploration into local vernacular housing trends and the history that shaped them. This variety, rather than signaling a lack of cohesion, represents a working-class neighborhood’s flexibility in response to local and national trends
Antimicrobial peptide class that forms discrete beta-barrel stable pores anchored by transmembrane helices
No full textPlease refer to the methods section within the associated publication.Bacteriocins are weapons of inter-bacterial warfare and belong to the larger group of antimicrobial peptides (AMPs), which are frequently proposed as alternatives to antibiotics. Many AMPs kill by destroying the target’s cytoplasmic membrane using short-lived membrane perturbation. Contrastingly, protein toxins form large pores by stably assembling in the target membrane. Here we describe an AMP family we termed TMcins (for transmembrane helix-containing bacteriocin), in which half of the AMP forms a transmembrane helix. This characteristic allows TMcin to assemble into stable and large oligomeric pores. The biosynthetic locus of TMcin, which was broadly active against Gram-positive bacteria, is distributed throughout two major bacterial phyla, yet bears no homology to previously reported bacteriocin biosynthetic gene clusters. Our discovery of an AMP class that achieves pore stability otherwise only found in protein toxins transforms our current understanding of AMP structure and function and underscores the continuing importance of phenotype-initiated investigations in uncovering wholly uncharacterized antimicrobials.This work was funded by the NIH Intramural Research Programs of the National Institute of Allergy and Infectious Diseases (project number ZIA AI000904 to M.O.), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (project number ZIA HD000072 to S.M.B.), the National Institute of Allergy and Infectious Disease BCBB Support Services Contract (HHSN316201300006W/75N93022F00001 to Guidehouse, Inc., to M.G.), the Canadian Institutes of Health Research (to N.C.J.S and D.P.T), the Digital Research Alliance of Canada (to D.P.T), the Canada Research Chairs program (to D.P.T) and the University of Maryland startup funds (to S.W.D.)
The Mechanicsburg Commercial Historic District: A Case Study of The National Register Amendment Process & Analysis
No full textThe Pennsylvania State Historic Preservation Office approved the Mechanicsburg Commercial Historic District, located in Cumberland County, Pennsylvania, for listing on the National Register of Historic Places in 1983. According to the nomination, the historic district consists of 102 buildings with a suggested period of significance beginning in 1800 and continuing through the present. The entire document is eight pages long with half dedicated to the physical description and historic narrative of the district. In my prior research, I looked into the Black history of Mechanicsburg and identified it as a crucial area missing from this National Register form. For this project, I will research the underrepresented histories of people who made significant contributions within this historic district and amend the record, which will be reviewed by Pennsylvania’s Historic Preservation Board and the National Park Service. In addition to the National Register amendment, I will write an analysis of the National Register amendment process, the difficulties therein, and potential areas for improvement
Fast Prediction of Full Quantum Dynamics with Deep Recurrent Neural Networks
No full textGemstone Team ASQNumerical simulations of interacting quantum systems are computationally very intensive, typically requiring resources that scale exponentially in the number of particles. A plausible approach to overcoming this unfavorable simulation time is to train deep neural networks over short timescales and use them to infer dynamics over much longer timescales. We demonstrate that such a speedup is possible using deep recurrent neural networks, including LSTM and Transformer-based networks, by predicting the quantum dynamics of multiple ’classic’ systems - the Ising, Heisenberg, and Hubbard models, with up to 9 spins, at most. We observe up to 3 orders of magnitude of data generation speedup for systems that can still be simulated with full system evolution. We observe this same performance - O(0.1) seconds - to generate data samples that cannot be generated with the resources we have available to use. Unique to our work, we predict the full wavefunction dynamics of the systems, which can then be used to calculate the
evolution of measurable and theoretical observables over time. We present sample predictions for our models and compare the efficacy of the different approaches with varying context-length
for prediction, spin count, and Hamiltonian parameters (mixing, interaction strength, etc.), at
best accurately predicting (< 10−6 mean square error - MSE) up to 90% of a single period with
10% of a period for context. We probe a number of frustrations, including square and triangular
interaction lattices, more complex next-nearest neighbor interactions, etc. to understand what
currently limits strong machine learning - ML - results in this space. Our results indicate that
the primary inhibitor to fast prediction at scale is the system scale, not the complexity of the dynamics. We anticipate that our work will provide insights towards extending the coherence time of quantum systems such as qubits and spins by determining the issues that stand in the way of network training and prediction on realistic Hamiltonians. We further believe that this work has immediate application in the simulation of large-scale neutral atom arrays, like Yt-171, under so-called Lieb-Robinson bounds
Deriving Vegetation Variables from Satellite Observations using a Data-driven Approach
No full textThis report was written at the completion of the 2025 CISESS Summer Internship Program.Land remote sensing techniques offer an unprecedented spatiotemporal coverage of key vegetative indicators in climatological modeling and understanding, such as fraction of vegetation cover (fCover). However, temporal inconsistencies and noise from snow cover, atmospheric contaminants, and viewing and illumination geometry hinder its applicability, necessitating validation using a more limited set of ground-based observations (GBOV). The research assesses three regression models: Cubist, XGBoost, and random forest for predicting ground-measured fCover using satellite-derived feature information. Ground measurements of fCover from 43 National Ecological Observatory Network (NEON) sites were processed at 20 m spatial resolution to provide labels for training, validation, and testing, which were then upscaled to 500 m to align with the high spatial resolution land surface reflectance data provided by the Visible Infrared Imaging Radiometer Suite (VIIRS) daily surface reflectance (VNP09GA) product. When evaluated against unseen data, the random forest regression model demonstrated the best agreement (R-squared = 0.912, MAE = 0.043), followed by the XGBoost regressor (R-squared = 0.910, MAE = 0.043) and the Cubist model (R-squared = 0.904, MAE = 0.047). Applying the random forest model on the 2023 VIIRS data for the East Coast produced estimates consistent with the expected annual phenological cycle. Limitations on the NEON site measurements may reduce the global representativeness and produce biases within regression models. Future work should focus on direct validation of the performance and representativeness using existing global products, such as GEOV3 and MODIS, and on the more globally representative BELMANIP2 sites.This study was supported by NOAA grant NA24NESX432C0001 (Cooperative Institute for Satellite Earth System Studies - CISESS) at the University of Maryland/ESSIC