5,054 research outputs found
MP-HULA
No full textDatacenter networks ofer a large degree of multipath in order to provide large bisectional bandwidth. The end-to-end performance is determined by the load-balancing strategy which needs to be designed to efectively manage congestion. Consequently, congestion aware load-balancing strategies such as CONGA or HULA have been designed. Recently, more and more applications that are hosted on cloud servers use multipath transport protocols such as MPTCP. However, in the presence of MPTCP, existing load-balancing schemes including ECMP, HULA or CONGA may lead to suboptimal forwarding decisions where multiple MPTCP subfows of one connection are pinned on the same bottleneck link. In this paper, we present MP-HULA, a transport layer multi-path aware load-balancing scheme using Programmable Data Planes. First, instead of tracking congestion information for the best path towards the destination, each MP-HULA switch tracks congestion information for the best-k paths to a destination through the neighbor switches. Second, we design MP-HULA using Programmable Data Planes, where each leaf switch can identify, using P4, which MPTCP subfow belongs to which connection. MP-HULA then load-balances diferent MPTCP subfows of a MPTCP connection on diferent next hops considering congestion state while aggregating bandwidth. Our evaluation shows that MP-HULA with MPTCP outperforms HULA in average flow completion time (2.1x at 50% load, 1.7x at 80% load).</p
Application of the IS-MP-IA model to the German economy and policy implications
No full textExtending the IS-MP-IA model developed by Romer (2000) and applying the GARCH (Engle, 1982, 2001) methodology, the author finds that equilibrium GDP in Germany is positively affected by stock market performance and real exchange rate appreciation, and negatively influenced by the expected inflation rate, the government deficit/GDP ratio, and the U.S. federal funds rate. The relatively low deficit/GDP ratio of 1.83% in 2003 indicates that its fiscal condition was healthy. However, some other EU members may need to exercise fiscal discipline. Because real appreciation has a positive impact on output, a stronger euro may not be a concern for Germany but may be worried by those EU member nations which depend upon exports to stimulate their economies.
Winning redefined, a new brand positioning for MP Motorsport
No full textMP Motorsport is a talent educating race team participating in the classes below the Formula 1. The problem with the lower classes and therefore MP Motorsport as well is that these classes don’t get much attention and exposure. As a company run by people with passion the branding and positioning is more something that naturally emerged from this passion rather than a series of conscious decisions. This makes it fragile. A more conscious strategy and vision on how MP Motorsport needs to be branded and positioned against their competition that is more than “look how cool racing is” is needed to become more attractive for sponsoring. This report describes the process from analysis to finding the right positioning and an advise for a new brand identity and an implementation plan to help becoming more attractive for sponsoring goals. The analysis done with employees showed a unique characteristic that will help the team to position themselves against the competition; the family atmosphere. This atmosphere is what also characterises their contradictory personality. This personality is on one hand leading and ambitious and on the other hand modest and involved. Competition and stakeholder analysis have been performed to find the right combination of being unique to the competition, desirable for the stakeholders (the fans and sponsors) and builds upon the core strengths of the company. Then using the brand key model, a positioning is made with the essence: Be your best self. The belief, “in everyone hides a talent” and values like “everyone is equal” and “together we succeed” combined with the essence and the positioning resulted in a brand story that describes the feeling MP Motorsport wants to communicate. To manifestate the brand, an advise has been set up for a tone of voice and tone of image. Wrapped in a concept called “winning redefined” this advise is part of the whole implementation plan that should lead the way for the team to implement the newly created brand in short and long term actions. In 3 phases MP Motorsport is advised to start with a clear brand introduction to the target groups Gen Z and potential sponsors. The next phase revolves around creating a community to attract both Gen Z talents and subsequently sponsors to interact with the team and each other. Finally the last phase is long term focused and aims for sustainable growth. In this phase the community is established and can expect various opportunities to discover and develop a whole range of talents, sponsors and gen z are connected to each other via MP Motorsport and the team is able to finance the lower classes without the pressure of the money drivers bring along.Strategic Product Desig
Derivations (MP) and Evaluations (OT)
No full textThe main claim of this paper is that the minimalist framework and optimality theory adopt more or less the same architecture of grammar: both assume that a generator defines a set S of potentially well-formed expressions that can be generated on the basis of a given input, and that there is an evaluator that selects the expressions from S that are actually grammatical in a given language L. The paper therefore proposes a model of grammar in which the strengths of the two frameworks are combined: more specifically, it is argued that the computational system of human language CHL from MP creates a set S of potentially well-formed expressions, and that these are subsequently evaluated in an optimality theoretic fashion.The definitive version of this paper is published in Linguistics in Potsdam 25 (2006).Broekhuis, H. (2006). Derivations (MP) and Evaluations (OT)*. In H. Broekhuis & R. Vogel (Eds), Linguistics in Potsdam 25. Optimality Theory and Minimalism: A possible Convergence? Potsdam : Universitätsverlag PotsdamISBN: 9783939469544 (published book)This research is supported by the Netherlands Organisation of Scientific Research (NWO), grant 276-70-00
MP-SMO: an algorithm for the VLSI implementation of the support vector machines training.
No full textMáquinas de aprendizagem, como Redes Neuronais Artificiais (ANNs), Redes Bayesianas, Máquinas de Vetores de Suporte (SVMs) e outras, são aplicadas em problemas de classificação de padrões. Devido ao baixo erro de teste, a SVM possui uma grande quantidade de aplicações, como no reconhecimento de imagens, seleção de genes, classificação de textos, robótica, reconhecimento de escrita a mão e outras. Dos algoritmos desenvolvidos para o treinamento da SVM, o Sequential Minimal Optimization (SMO) é um dos mais rápidos e o mais fácil de implementar em software. Devido a sua importância, várias otimizações para diminuir ainda mais o seu tempo de execução têm sido reportadas. A maioria das implementações do treinamento da SVM foram realizadas em software. Não obstante, a implementação em hardware é necessária em algumas aplicações com restrições: de área, e/ou de energia e/ou de tempo de treinamento, por exemplo, em algumas aplicações portáveis ou móveis. Nas implementações em hardware anteriores a este trabalho, o treinamento da SVM foi realizado com um conjunto de exemplos cuja quantidade é da ordem de somente dezenas, e unicamente uma delas usou o algoritmo SMO. Neste trabalho é apresentada uma modificação do algoritmo SMO, que denominamos algoritmo SMO de Múltiplos Pares (MP-SMO), para a aceleração do treinamento da SVM. A diminuição do tempo de treinamento é obtida realizando a otimização de um ou mais pares de coeficientes, chamados Multiplicadores de Lagrange, em cada iteração. De modo diferente, o algoritmo SMO original otimiza somente um par. O algoritmo MP-SMO apresenta as seguintes características: 1) a otimização de cada par de coeficientes é mantida simples usando a solução analítica do algoritmo SMO original. 2) as heurísticas para a seleção dos múltiplos pares a otimizar são adaptações das soluções anteriores para a seleção de um par por iteração. Testou-se o algoritmo otimizando até dois, três e quatro pares de coeficientes por iteração, e melhores resultados foram obtidos quando comparados com os do algoritmo SMO. Nos testes realizados com sete benchmarks, o tempo de treinamento diminuiu entre 22,5% e 42,8%. A diminuição do tempo de execução do algoritmo SMO em hardware é também abordada nesta dissertação. Os algoritmos SMO e MP-SMO foram completamente implementados em hardware dedicado para o benchmark Tic-tac-toe endgame. Este benchmark é composto por 958 exemplos, uma quantidade superior às usadas nas implementações anteriores. Com o algoritmo MP-SMO pretendeu-se reduzir o número de iterações, como na implementação em software, e poder incluir paralelismo na implementação em hardware. Para diminuir o tempo de execução de cada iteração, arquiteturas dos tipos pipeline e paralela foram usadas. Foram implementadas e testadas em um dispositivo do tipo FPGA (Field Programmable Gate Array) dezesseis diferentes arquiteturas no total, combinando ou não o algoritmo SMO ou o MP-SMO com pipelining e/ou paralelismo. O tempo de treinamento diminuiu no melhor caso para 1,8% do obtido com o algoritmo SMO implementado sem pipelining nem paralelismo, ou seja, diminuiu em mais de 50 vezes. Esta dissertação apresenta também a análise do custo em área e potência decorrente do aumento da velocidade de treinamento.Learning Machines, like Artificial Neural Networks (ANNs), Bayesian Networks, Support Vector Machines (SVMs) and others are applied in pattern classification problems. As the test error in SVM is small, it has several applications, such as image recognition, gene selection, text classification, robotics, handwritten recognition and others. Among the developed algorithms for the SVM training, the Sequential Minimal Optimization (SMO) is one of the fastest and the simplest to implement in software. Due to its importance, many improvements have been proposed in order to obtain even faster solutions than the original algorithm. Most of the SVM training implementations are in software. However, in some applications with restrictions of: area, and/or power and/or training time, a hardware implementation is necessary, for example, in some mobile or portable applications. In related previous works, the SVMs were trained in hardware using sets of only tens of examples, and in only one implementation the SMO algorithm was employed. In this work, a modified version of the SMO algorithm, named here the Multiple Pairs SMO (MP-SMO) algorithm, for the SVM training acceleration is presented. The training time reduction is obtained optimizing per iteration one or more pairs of coefficients known as Lagrange Multipliers, instead of only one pair as in the original SMO algorithm. The MP-SMO algorithm has the following features: 1) the optimization of each pair is as simple as in the original SMO algorithm because of the use of the same analytical method. 2) the solution for the pairs of coefficients selection can be chosen between two adapted heuristics for the SMO algorithm. The algorithm was tested optimizing up to two, three and four pairs of coefficients per iteration, and the training time was improved, when compared against the SMO algorithm. The tests for seven benchmarks showed an improvement that ranged from 22.5% to 42.8%. The reduction of the training time of the SMO algorithm executed in hardware is also treated in this dissertation. The algorithms SMO and MP-SMO were completely implemented in dedicated hardware for the Tic-tac-toe endgame benchmark. This benchmark is composed of 958 examples, a number greater than the used in the previous hardware implementations. The implementation of the MP-SMO algorithm is intended to reduce the number of iterations, as in the software implementation, and to include parallelism in the hardware implementation. In order to reduce the iteration execution time, the pipeline and parallel architectures were realized. Sixteen different architectures were implemented and tested on a Field Programmable Gate Array (FPGA) device, combining or not the SMO or MP-SMO algorithm with pipelining and/or parallelism. The training time was reduced to 1.8% of that obtained with the SMO algorithm without neither pipelining nor parallelism, that is, more than 50 times. This dissertation also presents an analysis of the area and power cost of the training speed increase
Stroom- en sedimentmeting Roompot-Veerse Dam (mp.8), 15 en 22 januari 1996
No full textStroom- en sedimentmeting Roompot-Veerse Dam (mp.8) 15 en 22 januari 1996.Deltawerken, Oosterscheld
Removal of domain D2 or D3 of the human urokinase receptor does not affect ligand affinity.
No full textA novel approach to MP-PIC: Continuum particle model for dense particle flows in fluidized beds
No full textA novel approach to Multiphase-Particle-in-Cell (MP-PIC), called Continuum Particle Model (CPM), is developed for dense gas-particle flows. CPM has high computational speed, comparable to that of MP-PIC, but a robustness and accuracy closer to that of a Discrete Element Model (DEM). The gas phase is treated as a continuum phase and particles are tracked discretely, but particle collisions are modelled by considering the divergence of the continuum particle stress tensor. Details on efficient solution to the model are presented. For comparison, a parametric study is performed for quasi-2D fluidized beds. Comparison of CFD-CPM is made with MP-PIC and CFD-DEM. The particle stress models by Harris and Crighton, and by Srivastava and Sundaresan are tested in our CFD-CPM. Results from CFD-CPM based on the Srivastava and Sundaresan particle stress model show good agreement with CFD-DEM results. We validate our model by comparison with experimental benchmark results from Gopalan et. al. (2016).Complex Fluid Processin
Comparison of Two HCV Confirmatory Tests: MP Diagnostics HCV Blot 3.0 and CHIRON RIBA HCV 3.0
No full textThe aim of this study is to evaluate the MP Diagnostics HCV Blot 3.0 (MP Blot) as a supplemental confirmatory assay for the presence of antibodies to HCV in comparison with CHIRON RIBA HCV 3.0 (RIBA). After the discontinuation of RIBA, other antibody test should be validated for distinguish between resolved HCV infection and biologic false positivity for HCV antibody in persons in whom HCV RNA is not detected. A total of 176 samples from routine patients screened in our laboratory were selected for this retrospective study. MP Blot significantly reduces indeterminate results due to reduced sensitivity to NS3 and to the differences of manufacturers’ interpretative criteria: only one core band with 2+ or greater reactivity is considered positive on MP Blot, but indeterminate on RIBA. We observed risk of false negative results due to this reduced sensibility to NS3, requiring PCR and sequential serological control in patients with suspicion of early HCV seroconversion.info:eu-repo/semantics/publishe
Genomewide linkage scan of schizophrenia in a large multicenter pedigree sample using single nucleotide polymorphisms
No full textA genomewide linkage scan was carried out in eight clinical samples of informative schizophrenia families. After all quality control checks, the analysis of 707 European-ancestry families included 1615 affected and 1602 unaffected genotyped individuals, and the analysis of all 807 families included 1900 affected and 1839 unaffected individuals. Multipoint linkage analysis with correction for marker-marker linkage disequilibrium was carried out with 5861 single nucleotide polymorphisms (SNPs; Illumina version 4.0 linkage map). Suggestive evidence for linkage ( European families) was observed on chromosomes 8p21, 8q24.1, 9q34 and 12q24.1 in nonparametric and/or parametric analyses. In a logistic regression allele-sharing analysis of linkage allowing for intersite heterogeneity, genomewide significant evidence for linkage was observed on chromosome 10p12. Significant heterogeneity was also observed on chromosome 22q11.1. Evidence for linkage across family sets and analyses was most consistent on chromosome 8p21, with a one-LOD support interval that does not include the candidate gene NRG1, suggesting that one or more other susceptibility loci might exist in the region. In this era of genomewide association and deep resequencing studies, consensus linkage regions deserve continued attention, given that linkage signals can be produced by many types of genomic variation, including any combination of multiple common or rare SNPs or copy number variants in a region. Molecular Psychiatry (2009) 14, 786-795; doi:10.1038/mp.2009.11; published online 17 February 2009</p
- …
