University of Maryland, Baltimore County
Not a member yet
17643 research outputs found
Sort by
Experimental evaluation and implementation of Spread Identity Framework.
No full textWe provide a novel implementation of the Spread Identity (SI) framework and experimentally evaluate it on an organizational level. Unlike previous implementations of SI, ours builds on RFC2663 and RFC2694, which describes approaches for application-level gateways for domain name systems and twice network address translation, respectively. We implement a minimal system suitable for a medium-sized organization and measure connection establishment time, network address translation (NAT) table size, name resolution time, packet forwarding rate, and memory requirements. We find that our implementation consumes similar resources to those of a typical single NAT system. Dhananjay Phatak, et al. first proposed SI in 2005. The central idea is to spread identity of a host across multiple network addresses through dynamic mapping of IP addresses to host names, thereby obfuscating the host's true network identity to some degree. One application is to mitigate IP packet floods. Using the ns-2 network simulator, in 2010 Sonawane experimentally demonstrated the effectiveness of SI as a defense against packet flood attacks. By contrast, our work evaluates the performance of our new implementation of SI on a real network. Using the Deterlab test network, we experimentally compare our SI implementation with a single NAT subsystem, with both systems using NetFilter for NATting. We generate 100,000 simultaneous connections to test the baseline NAT and SI systems (a typical load for medium-sized organizations). Except for name resolution, both systems perform similarly on all measured parameters. In particular, we find there is no difference between end-to-end connection establishment time, up to 90 us difference in gateway packet translation time. We also find that NAT table size cannot be considered in isolation: NAT table entries are a part of connection tracking entries which require 304 byte entries per connection for both our implemented baseline and SI systems. In SI, name resolution takes up to 1500ms in the worst case, compared to 50ms in the baseline system. From these results, we conclude that our system is technically feasible and results in similar load patterns on the gateway to those from a typical single NAT system
Cognitive and Behavioral Development Among Children Prenatally Exposed to Substances: An Examination of Models of Risk
No full textThe combined effects of drug exposure and environmental adversity may have important implications for several developmental outcomes. The existent literature presents mixed findings regarding the direct effect of cumulative risk among drug exposed samples on cognitive and behavioral outcomes. The present research was divided into two studies to examine the nature of the relation between cumulative risk during the first year of life and cognitive and behavioral outcomes at 60 months. Both studies used a sample of 132 primarily low-income African American children exposed to cocaine (30.8%), heroin (14.0%), or both (55.1%). The first study examined cumulative risk using four models: cumulative risk, threshold effects, specific indices, and unique effects. There were no direct linear or nonlinear relations between cumulative risk with cognitive or behavioral outcomes. There was a direct association between psychosocial risk (specifically negative life events) with both internalizing and externalizing behaviors. None of the models of risk were associated with intelligence. The second study examined process variables (mental development scores, harsh caregiving, and caregiver change) to determine whether proximal risk factors may explain the relation between cumulative risk and cognitive and behavioral development. The results of the second study suggested that mental development scores at 12, 18, 24, and 30 months were predictive of intelligence at 60 months, but not behavior. Harsh caregiving was directly associated with internalizing behaviors, and moderated the relation between cumulative risk and internalizing behaviors. Finally, the number of caregiver changes from birth-60 months was associated with greater attention problems at 60 months. Overall, the findings from the present study provide little support for an association between cumulative risk with developmental outcomes within a drug exposed sample. The overall cognitive scores were very low and not associated with any risk factors, suggesting that the present sample was relatively low functioning, cognitively. Psychosocial risk was associated with behavior problems, especially caregiver negative life events and harsh caregiving. These findings suggest that the caregiving environment represents an important context for behavioral development. Implications for theory, practice, and policy are discussed
MULTI-SCALE NUMERICAL STUDY OF NANOPARTICLE TRANSPORT IN DEFORMABLE TISSUES DURING AN INFUSION PROCESS
No full textNanoparticles have attracted growing research interest in novel hyperthermia treatments of cancers due to their ability to generate impressive levels of heat when excited by an external alternating magnetic field. Intratumoral injection by means of needle is an important method to deliver nanoparticles in tumors, and the resulting distribution of the particles is a critical factor determining the distribution of temperature elevations. In order to maximize the outcome of magnetic nanoparticle hyperthermia treatment, it is essential to acquire knowledge and understanding of the transport of nanoparticles and nanofluid in biological tissues during and after an infusion process. A multiscale model is developed in this dissertation that consists of a macroscale model for nanofluid flow, tissue deformation, nanoparticle convection and diffusion, as well as nanoparticle deposition in porous tissues, and a particle trajectory tracking model that considers particle-surface interactions and the resulting effect of particle deposition. The integration of these two models at different scales enables the study of the behavior of nanoparticle deposition on the cellular structure, the formation of backflow and the variation of tissue porosity induced by an infusion, the spreading shape of nanofluid, and the transport mechanism of nanoparticles in a tumor. The particle deposition rate coefficient, the shapes of nanofluid distribution volume, the length of backflow, and the distribution of nanoparticle concentration are predicted for a wide range of injection conditions. The results show that tissue deformation and particle binding to the cell surface are leading mechanisms that substantially affect the particle concentration distribution and the nanofluid spreading shape after an intratumoral injection. Infusion-induced tissue deformation results in backflow that causes nanofluid to flow back and infiltrate along the needles track, thus leading to an irregular shape of the nanofluid spreading. Tissue deformation also alters the pore size near the injection site, and thereby facilitates particle penetration through the tissue. Particle binding to the cellular structure, which reduces the particle concentration in the fluid phase, is found to be the leading mechanism causing the non-uniform distribution of the nanoparticles after an injection. The rate of particle binding to the cellular structure can be reduced by increasing the local interstitial fluid velocity, tissue porosity, and surface charge of the particles. Simulations are also performed under various infusion conditions, and the effects of injection rate, needle sizes, tissue elastic properties, and injection amounts on the spatial distribution of nanoparticles in tumors are quantified
EMPIRICAL STUDY ON TECHNICAL DEBT AS VIEWED BY SOFTWARE PRACTITIONERS
No full textThe Technical Debt metaphor is gaining recognition in the software development community, not only to communicate the debt to the non-technical stakeholders but also to maintain the long term health of the software project. The idea is that software practitioners accept compromises in the project in some magnitude to meet demand in other dimensions. The compromises in the system incur a debt which has to be repaid at some point for the long term health of the project. If the debt remains unpaid, it could incur more additional costs, in terms of software that is harder to change and more error prone. This pain is most evident to the software practitioners, as they are the ones' who have to deal with these problems at the code level. We have conducted an interview study of software practitioners to analyze and understand how they face this debt and how they manage it. The result of this study identifies the characteristics, causes, and approaches for management of technical debt. As a result of analysis of the data for this study we also present several models of the cause and effect of different types of technical debt. These models also provide an understanding upon which approaches to track, measure, quantify and monitor technical debt can be developed
Finding a Temporal Comparison Function for the Metacognitive Loop
No full textThe field of Artificial Intelligence has seen steady advances in cognitive systems. However, many of these systems perform poorly when faced with situations outside of their training. Since the real world is dynamic, this brittleness is a major problem in the field today. Adding metacognition to such systems can improve their operation in the face of perturbations found in dynamic environments. I developed a six-level taxonomy that divided metacognitive systems according to their capabilities. This ranged from Level 0: Bereft that offered no metacognitive assistance, to Level 5: Anticipating that would make suggestions before the agent needed assistance. Using this taxonomy, I deconstructed an existing metacognitive system (MCL) so that it could operate at any of the four lower levels of the taxonomy. I added, to the Level 3: Temporal version of MCL, a number of comparison functions designed to determine if the problem the agent is currently facing is similar to any previous problem. I created functions to both establish baseline operation performance and to try to optimize the agent's performance in a perturbed environment. I extended an existing Mars Rover domain simulation to add more problem and recovery options. In this environment, I tested the temporal comparison functions to determine how well they were able to distinguish on type of perturbation from another and how much they helped (or hindered) the agents subjected to those perturbations. Several comparison functions were able to aid the Rover in completing its tasks, although none were perfect
Digital Storytelling in ESL Instruction: Identity Negotiation through a Pedagogy of Multiliteracies
No full textThis qualitative descriptive exploratory study investigated how a pedagogy of multiliteracies can be introduced to the ESL curriculum using digital stories and explored the ways ESL learners negotiate their identities through the multimodality of this narrative genre. The study was based on the premise that in language education it is crucial to account for multimodality of discourses and bring students' diverse lifeworlds and experiences into the classroom thus extending an understanding of literacy to multimodal communication and inviting ESL learners to explore their multilayered and dynamic identities. The study elicited data through focused participant observations, content analysis of students' essays, weekly journals, and final semi-structured interviews, and through discourse analysis of the students' drafts of verbal narratives for digital stories, digital storyboards, and final digital storytelling projects. The findings revealed the presence of situated practice; overt instruction; critical framing; and, in a more limited way, transformed practice. While situated practice evolved with the students narrating about their families, life-changing events, and important cultural practices, overt instruction included explicit and systematic instruction and scaffolding that fostered students' multimodal meaning making through story writing and production. Combined with situated practice it resulted in critical framing when students reflected on their progress as language learners and analyzed multimodal cultural representations in their digital stories. Some nascent examples of transformed practice were evident when students suggested how the projects had influenced their overall approaches to learning and understanding of meaning making. Students negotiated their identities through the process and product of digital stories. The study revealed that none of the participants articulated as central an ESL student identity indicating that this social role was not particularly significant at the time of the digital story project. Instead, the students were storytellers and producers, mothers and sons, daughters and sisters, granddaughters and world travelers, friends, women and men, and people with unique cultural backgrounds and social experiences. And since the students were in constant communication with each other, personal stories became experiences that moved other students and influenced their understanding of cultural diversity. This collaborative process created a multicultural classroom community of practice conducive to the recognition of diverse identities and social roles
Grrl Parts (2011)
No full textPromotional materials for Grrl Parts, performed in Spring 2011. The three plays performed were Snip, Runaways, and What is the Custom of Your Grief? Includes a playbill, theater program, and publicity photographs.Additional photographs can be found in UARC 2011-021March 2-6, 201
[EB 2011, Student/Postdoc Poster Reception]
No full text[Title supplied by cataloger]Jennifer Keller explains her research to ACYA poster judges at the Student/Postdoc Poster Reception, at the 2011 Experimental Biology conference