Overlay Network Creation and Maintenance with Selfish Users

Overlay networks have been used for adding and enhancing functionality to the end-users without requiring modifications in the Internet core mechanisms. Overlay networks have been used for a variety of popular applications including routing, file sharing, content distribution, and server deployment. Previous work has focused on devising practical neighbor selection heuristics under the assumption that users conform to a specific wiring protocol. This is not a valid assumption in highly decentralized systems like overlay networks. Overlay users may act selfishly and deviate from the default wiring protocols by utilizing knowledge they have about the network when selecting neighbors to improve the performance they receive from the overlay. This thesis goes against the conventional thinking that overlay users conform to a specific protocol. The contributions of this thesis are threefold. It provides a systematic evaluation of the design space of selfish neighbor selection strategies in real overlays, evaluates the performance of overlay networks that consist of users that select their neighbors selfishly, and examines the implications of selfish neighbor and server selection to overlay protocol design and service provisioning respectively. This thesis develops a game-theoretic framework that provides a unified approach to modeling Selfish Neighbor Selection (SNS) wiring procedures on behalf of selfish users. The model is general, and takes into consideration costs reflecting network latency and user preference profiles, the inherent directionality in overlay maintenance protocols, and connectivity constraints imposed on the system designer. Within this framework the notion of user’s "best response" wiring strategy is formalized as a k-median problem on asymmetric distance and is used to obtain overlay structures in which no node can re-wire to improve the performance it receives from the overlay. Evaluation results presented in this thesis indicate that selfish users can reap substantial performance benefits when connecting to overlay networks composed of non-selfish users. In addition, in overlays that are dominated by selfish users, the resulting stable wirings are optimized to such great extent that even non-selfish newcomers can extract near-optimal performance through naïve wiring strategies. To capitalize on the performance advantages of optimal neighbor selection strategies and the emergent global wirings that result, this thesis presents EGOIST: an SNS-inspired overlay network creation and maintenance routing system. Through an extensive measurement study on the deployed prototype, results presented in this thesis show that EGOIST’s neighbor selection primitives outperform existing heuristics on a variety of performance metrics, including delay, available bandwidth, and node utilization. Moreover, these results demonstrate that EGOIST is competitive with an optimal but unscalable full-mesh approach, remains highly effective under significant churn, is robust to cheating, and incurs minimal overheads. This thesis also studies selfish neighbor selection strategies for swarming applications. The main focus is on n-way broadcast applications where each of n overlay user wants to push its own distinct file to all other destinations as well as download their respective data files. Results presented in this thesis demonstrate that the performance of our swarming protocol for n-way broadcast on top of overlays of selfish users is far superior than the performance on top of existing overlays. In the context of service provisioning, this thesis examines the use of distributed approaches that enable a provider to determine the number and location of servers for optimal delivery of content or services to its selfish end-users. To leverage recent advances in virtualization technologies, this thesis develops and evaluates a distributed protocol to migrate servers based on end-users demand and only on local topological knowledge. Results under a range of network topologies and workloads suggest that the performance of the distributed deployment is comparable to that of the optimal but unscalable centralized deployment.

Biased ART: A Neural Architecture that Shifts Attention Toward Previously Disregarded Features Following an Incorrect Prediction

Memories in Adaptive Resonance Theory (ART) networks are based on matched patterns that focus attention on those portions of bottom-up inputs that match active top-down expectations. While this learning strategy has proved successful for both brain models and applications, computational examples show that attention to early critical features may later distort memory representations during online fast learning. For supervised learning, biased ARTMAP (bARTMAP) solves the problem of over-emphasis on early critical features by directing attention away from previously attended features after the system makes a predictive error. Small-scale, hand-computed analog and binary examples illustrate key model dynamics. Twodimensional simulation examples demonstrate the evolution of bARTMAP memories as they are learned online. Benchmark simulations show that featural biasing also improves performance on large-scale examples. One example, which predicts movie genres and is based, in part, on the Netflix Prize database, was developed for this project. Both first principles and consistent performance improvements on all simulation studies suggest that featural biasing should be incorporated by default in all ARTMAP systems. Benchmark datasets and bARTMAP code are available from the CNS Technology Lab Website: http://techlab.bu.edu/bART/.

Novel all-optical on-off-keyed to alternate-mark-inversion converter

We numerically investigate a novel 40 Gbps OOK to AMI all-optical modulation format converter employing an SOA-based Mach-Zehnder interferometer.  We demonstrate operation with a 27-1 PRBS and explain the phase modulation's relationship with patterning.

Use of semiconductor optical amplifiers in signal processing applications

We describe a 42.6 Gbit/s all-optical pattern recognition system which uses semiconductor optical amplifiers (SOAs). A circuit with three SOA-based logic gates is used to identify the presence of specific port numbers in an optical packet header.

WIMU instrumentation of assassin trainer & skeleton sled

Skeleton is a high‐speed Winter Olympic sport performed on the same twisting, downhill ice tracks used for Bobsleigh & Luge. The single rider sprints and pushes their sled for 20‐30m on a level start section before loading and going through a twisting course of over 1km, at speeds up to 140km/h, experiencing up to 5g. In competition, the top athletes can be within a fraction of a second of each other. The initial short pushing period is believed to be critical to overall performance but it is not well understood. A collaborative project between University of Bath, UK Sport and Tyndall National Institute is instrumenting skeleton athletes, training equipment and test tracks with Tyndall’s Wireless Inertial Measurement Unit technology in order to investigate and improve understanding of this phase of a skeleton run. It is hoped this will lead to improved training regimes and better performance of such elite, Olympic level athletes. This work presents an initial look at the system as implemented and data recorded.

An examination of the construction of meanings for obesity in Ireland

Obesity has been defined as a consequence of energy imbalance, where energy intake exceeds energy expenditure and results in a build-up of adipose tissue. However, this scientific definition masks the complicated social meanings associated with the condition. This research investigated the construction of meaning around obesity at various levels of inquiry to inform how obesity is portrayed and understood in Ireland. A multi-paradigmatic approach was adopted, drawing on theory and methods from psychology and sociology and an analytical framework combining the Common Sense Model and framing theory was employed. In order to examine the exo-level meanings of obesity, content analysis was performed on two media data sets (n=479, n=346) and a thematic analysis was also performed on the multiple newspaper sample (n=346). At the micro-level, obesity discourses were investigated via the thematic analysis of comments sampled from an online message board. Finally, an online survey assessed individual-level beliefs and understandings of obesity. The media analysis revealed that individual blame for obesity was pervasive and the behavioural frame was dominant. A significant increase in attention to obesity over time was observed, manifestations of weight stigma were common, and there was an emotive discourse of blame directed towards the parents of obese children. The micro-level analysis provided insight into the weight-based stigma in society and a clear set of negative ‘default’ judgements accompanied the obese label. The survey analysis confirmed that the behavioural frame was the dominant means of understanding obesity. One of the strengths of this thesis is the link created between framing and the Common Sense Model in the development of an analytical framework for application in the examination of health/illness representations. This approach helped to ascertain the extent of the pervasive biomedical and individual blame discourse on obesity, which establishes the basis for the stigmatisation of obese persons.

The mobile cloud - more than a cloud

Nearly one billion smart mobile devices are now used for a growing number of tasks, such as browsing the web and accessing online services. In many communities, such devices are becoming the platform of choice for tasks traditionally carried out on a personal computer. However, despite the advances, these devices are still lacking in resources compared to their traditional desktop counterparts. Mobile cloud computing is seen as a new paradigm that can address the resource shortcomings in these devices with the plentiful computing resources of the cloud. This can enable the mobile device to be used for a large range of new applications hosted in the cloud that are too resource demanding to run locally. Bringing these two technologies together presents various difficulties. In this paper, we examine the advantages of the mobile cloud and the new approaches to applications it enables. We present our own solution to create a positive user experience for such applications and describe how it enables these applications.

High-field FMRI reveals brain activation patterns underlying saccade execution in the human superior colliculus.

BACKGROUND: The superior colliculus (SC) has been shown to play a crucial role in the initiation and coordination of eye- and head-movements. The knowledge about the function of this structure is mainly based on single-unit recordings in animals with relatively few neuroimaging studies investigating eye-movement related brain activity in humans. METHODOLOGY/PRINCIPAL FINDINGS: The present study employed high-field (7 Tesla) functional magnetic resonance imaging (fMRI) to investigate SC responses during endogenously cued saccades in humans. In response to centrally presented instructional cues, subjects either performed saccades away from (centrifugal) or towards (centripetal) the center of straight gaze or maintained fixation at the center position. Compared to central fixation, the execution of saccades elicited hemodynamic activity within a network of cortical and subcortical areas that included the SC, lateral geniculate nucleus (LGN), occipital cortex, striatum, and the pulvinar. CONCLUSIONS/SIGNIFICANCE: Activity in the SC was enhanced contralateral to the direction of the saccade (i.e., greater activity in the right as compared to left SC during leftward saccades and vice versa) during both centrifugal and centripetal saccades, thereby demonstrating that the contralateral predominance for saccade execution that has been shown to exist in animals is also present in the human SC. In addition, centrifugal saccades elicited greater activity in the SC than did centripetal saccades, while also being accompanied by an enhanced deactivation within the prefrontal default-mode network. This pattern of brain activity might reflect the reduced processing effort required to move the eyes toward as compared to away from the center of straight gaze, a position that might serve as a spatial baseline in which the retinotopic and craniotopic reference frames are aligned.

Bayesian Gaussian Copula Factor Models for Mixed Data.

Gaussian factor models have proven widely useful for parsimoniously characterizing dependence in multivariate data. There is a rich literature on their extension to mixed categorical and continuous variables, using latent Gaussian variables or through generalized latent trait models acommodating measurements in the exponential family. However, when generalizing to non-Gaussian measured variables the latent variables typically influence both the dependence structure and the form of the marginal distributions, complicating interpretation and introducing artifacts. To address this problem we propose a novel class of Bayesian Gaussian copula factor models which decouple the latent factors from the marginal distributions. A semiparametric specification for the marginals based on the extended rank likelihood yields straightforward implementation and substantial computational gains. We provide new theoretical and empirical justifications for using this likelihood in Bayesian inference. We propose new default priors for the factor loadings and develop efficient parameter-expanded Gibbs sampling for posterior computation. The methods are evaluated through simulations and applied to a dataset in political science. The models in this paper are implemented in the R package bfa.

Neural networks supporting autobiographical memory retrieval in posttraumatic stress disorder.

Posttraumatic stress disorder (PTSD) affects the functional recruitment and connectivity between neural regions during autobiographical memory (AM) retrieval that overlap with default and control networks. Whether such univariate changes relate to potential differences in the contributions of the large-scale neural networks supporting cognition in PTSD is unknown. In the present functional MRI study, we employed independent-component analysis to examine the influence of the engagement of neural networks during the recall of personal memories in a PTSD group (15 participants) as compared to non-trauma-exposed healthy controls (14 participants). We found that the PTSD group recruited similar neural networks when compared to the controls during AM recall, including default-network subsystems and control networks, but group differences emerged in the spatial and temporal characteristics of these networks. First, we found spatial differences in the contributions of the anterior and posterior midline across the networks, and of the amygdala in particular, for the medial temporal subsystem of the default network. Second, we found temporal differences within the medial prefrontal subsystem of the default network, with less temporal coupling of this network during AM retrieval in PTSD relative to controls. These findings suggest that the spatial and temporal characteristics of the default and control networks potentially differ in a PTSD group versus healthy controls and contribute to altered recall of personal memory.

Dynamic neural networks supporting memory retrieval.

How do separate neural networks interact to support complex cognitive processes such as remembrance of the personal past? Autobiographical memory (AM) retrieval recruits a consistent pattern of activation that potentially comprises multiple neural networks. However, it is unclear how such large-scale neural networks interact and are modulated by properties of the memory retrieval process. In the present functional MRI (fMRI) study, we combined independent component analysis (ICA) and dynamic causal modeling (DCM) to understand the neural networks supporting AM retrieval. ICA revealed four task-related components consistent with the previous literature: 1) medial prefrontal cortex (PFC) network, associated with self-referential processes, 2) medial temporal lobe (MTL) network, associated with memory, 3) frontoparietal network, associated with strategic search, and 4) cingulooperculum network, associated with goal maintenance. DCM analysis revealed that the medial PFC network drove activation within the system, consistent with the importance of this network to AM retrieval. Additionally, memory accessibility and recollection uniquely altered connectivity between these neural networks. Recollection modulated the influence of the medial PFC on the MTL network during elaboration, suggesting that greater connectivity among subsystems of the default network supports greater re-experience. In contrast, memory accessibility modulated the influence of frontoparietal and MTL networks on the medial PFC network, suggesting that ease of retrieval involves greater fluency among the multiple networks contributing to AM. These results show the integration between neural networks supporting AM retrieval and the modulation of network connectivity by behavior.

THE ART SONG OF SOUTH AMERICA: AN EXPLORATION THROUGH PERFORMANCE

The repertoire included in this dissertation was presented over the course of three recitals, The Songs of Argentina, The Songs of Brazil, Chile and Venezuela, and The Songs of Perú and Colombia. Each recital was supplemented by written program notes and English translations of the Spanish, Portuguese and Quechua texts. The selections presented in this study was chosen in an effort to pair the works of internationally renowned composers like Argentine composers Alberto Ginastera and Carlos Guastavino, and Brazilian composer Heitor Villa-Lobos, with those of lesser-known composers, including Venezuelan composer Juan Bautista Plaza, Peruvian composers Edgar Valcárcel, Theodoro Valcárcel, and Rosa Mercedes Ayarza de Morales, and Colombian composer Jaime Léon. Each composer represents a milestone in the development of art song composition in South America. All three recitals were recorded and are available on compact discs in the Digital Repository at the University of Maryland (DRUM). This dissertation was completed in May, 2011.

CANCIONES ARGENTINAS (SONGS OFARGENTINA): A RECORDING PROJECT

This Dissertation Project comprises recordings of Argentine art songs. The discs are approximately 40-60 minutes in length and consist of songs from the traditional art-song repertoire for voice and piano. This project is particularly appropriate because of the very limited number of recordings of Argentine songs, which are notable both not only for their high quality but for their accessibility of performance for voice teachers, students, and professional singers alike. Art songs in the Spanish language are a welcome resource, and the poetry included in this project is of an outstanding quality. Some of the poets set to music are Gabriela Mistral (a poet laureate of Chile and the first Latin American woman to win the Nobel Prize for Literature), Pablo Neruda (also a Nobel laureate), Luis Cernuda, and Leon Benar6s. The lyrics of some songs are based on traditional sources, and the melodies and rhythms of all are representative of South American-indigenous and European­ immigrant cultures. The composers represented here will be familiar to some listeners but more than likely unfamiliar to most. Yet Alberto Ginastera (1916-1983) is considered to be the greatest of all Argentine composers. Alberto Williams (1862-1952) is known as the father of the Nationalist School of composition in Argentina, and Carlos Lopez Buchardo (1881-1963) is a most influential composer and pedagogue after whom the national Conservatory of Music in Buenos Aries is named. Two composers who remain relatively unknown outside of South America, Abraham Jurafsky (1906- 1993) and Julio Perceval (1903-1963) are also represented in this project. A complete compact disc is devoted to the works of Carlos Guastavino. Known as the "Argentine Schubert", Guastavino has over 250 songs to his credit. Chiefly a composer for piano and voice, his recent death (October 2000) makes a recording of his works especially appropriate. This project also includes a written component, a supportive dissertation briefly describing the history of the Argentine art song and the lives and influences of the composers and poets represented in the studio recordings. The CD recordings are held in the Michelle Smith Performing Arts Library at the University of Maryland.

A SURVEY OF FRENCH KEYBOARD MUSIC FROM 1880 TO 1950

French music flourished from the last quarter of the nineteenth century into the early twentieth century, especially in the genres of opera and orchestral music. Although French keyboard music enjoyed less popularity, being overshadowed by these predominant genres, prominent impressionist figures Claude Debussy (1862-1918) and Maurice Ravel (1875-1937) . . brought its revival to the French music. Scholars consider Debussy to be a frontrunner of Impressionism, and his influence had a major impact on subsequent composers. As a result of his popularity, other significant works by French composers seem to be overlooked by pianists and audiences and are not as often performed. Because keyboard works by Debussy and Ravel are a popular performance choice among pianists, I was eager to examine music by other French composers. Through my resea,rch, I found many great works that warranted further study and deserve a place in the keyboard repertoire. This recording project contains works by lesser-known French composers written between the years of 1880 and 1950, namely Emmanuel Chabrier (1841-1894), Gabriel-Urbain Faure (1845-1924), Charles Koechlin (1867-1950), Albert Roussel (1869-1937), Erik Satie (1866-1925), Francis Poulenc (1899-1963), Darius Milhaud (1892-1974), Robert Casadesus (1899-1972) and Henri Dutilleux (b.1916). Since piano repertoire is abundant, it is sometimes difficult to create a performance program. Therefore, it frequently becomes the default to choosing familiar repertoire rather than using the opportunity to expand the repertoire. As a pianist, I feel responsible to search for hidden musical treasures with which pianists and audiences alike are not so well acquainted. This recording project explores nine lesser-known French compositions written between 1880 and 1950. I expect this to be an opportunity to introduce both pianists and audiences to outstanding but unfamiliar works by French composers. This dissertation was recorded on two compact discs in Dekelboum Concert Hall at Clarice Smith Performing Arts Center of the University of Maryland. The recordings are archived in the University of Maryland Library.

Heteronormativity and practitioner-patient interaction.

Heteronormativity is the presumption of heterosexuality as the default sexual orientation and can result in discrimination against the lesbian, gay, and bisexual (LGB) population. This study serves as one of the first experimental studies to examine heteronormative perceptions in communication and their effects on practitioner-patient relationships. LGB participants were randomly assigned to read either heteronormative or non-heteronormative vignettes of a doctor-patient interaction. They then indicated how much health-relevant information they would disclose to the doctor in the vignette and their level of trust in the doctor. In the heteronormative condition, participants were less likely to disclose health-relevant information to the doctor in the vignette and were less trustful of the doctor as compared to those in the non-heteronormative condition. These results have important health implications, as lack of disclosure and trust may prevent people from getting needed care and prevent doctors from giving the best health advice possible. The results of this study provide further evidence that there is a need for more education for all health care professionals to feel comfortable while respectfully communicating with and treating patients who do not identify as heterosexual in order to ensure the best health care experience.