University of Maryland, Baltimore County

About: University of Maryland, Baltimore County is a education organization based out in Baltimore, Maryland, United States. It is known for research contribution in the topics: Population & Galaxy. The organization has 8749 authors who have published 20843 publications receiving 795706 citations. The organization is also known as: UMBC.

Electronic survey methodology: A case study in reaching hard-to-involve internet users.

Abstract: Using the Internet to conduct quantitative research presents challenges not found in conventional research. Paper-based survey quality criteria cannot be completely adapted to electronic formats. Electronic surveys have distinctive technological, demographic, and response characteristics that affect their design, use, and implementation. Survey design, participant privacy and confidentiality, sampling and subject solicitation, distribution methods and response rates, and survey piloting are critical methodological components that must be addressed. In this article, quality criteria for electronic survey design and use based on an investigation of recent electronic survey literature are presented. The application of these criteria to reach a hard-to-involve online population-nonpublic participants of online communities (also known as "lurkers")-and survey them on their community participation, a topic not salient to the purpose of their online communities is demonstrated in a case study. The results show t...

The genome sequence of Bacillus anthracis Ames and comparison to closely related bacteria

Abstract: Bacillus anthracis is an endospore-forming bacterium that causes inhalational anthrax. Key virulence genes are found on plasmids (extra-chromosomal, circular, double-stranded DNA molecules) pXO1 (ref. 2) and pXO2 (ref. 3). To identify additional genes that might contribute to virulence, we analysed the complete sequence of the chromosome of B. anthracis Ames (about 5.23 megabases). We found several chromosomally encoded proteins that may contribute to pathogenicity--including haemolysins, phospholipases and iron acquisition functions--and identified numerous surface proteins that might be important targets for vaccines and drugs. Almost all these putative chromosomal virulence and surface proteins have homologues in Bacillus cereus, highlighting the similarity of B. anthracis to near-neighbours that are not associated with anthrax. By performing a comparative genome hybridization of 19 B. cereus and Bacillus thuringiensis strains against a B. anthracis DNA microarray, we confirmed the general similarity of chromosomal genes among this group of close relatives. However, we found that the gene sequences of pXO1 and pXO2 were more variable between strains, suggesting plasmid mobility in the group. The complete sequence of B. anthracis is a step towards a better understanding of anthrax pathogenesis.

Can You See Me Now? Audience and Disclosure Regulation in Online Social Network Sites

Abstract: The prevailing paradigm in Internet privacy literature, treating privacy within a context merely of rights and violations, is inadequate for studying the Internet as a social realm. Following Goffm...

Super-Convergence: Very Fast Training of Residual Networks Using Large Learning Rates

Abstract: In this paper, we show a phenomenon, which we named ``super-convergence'', where residual networks can be trained using an order of magnitude fewer iterations than is used with standard training methods. The existence of super-convergence is relevant to understanding why deep networks generalize well. One of the key elements of super-convergence is training with cyclical learning rates and a large maximum learning rate. Furthermore, we present evidence that training with large learning rates improves performance by regularizing the network. In addition, we show that super-convergence provides a greater boost in performance relative to standard training when the amount of labeled training data is limited. We also derive a simplification of the Hessian Free optimization method to compute an estimate of the optimal learning rate. The architectures to replicate this work will be made available upon publication.