A pipelined architecture for distributed text query evaluation

Two principal query-evaluation methodologies have been described for cluster-based implementation of distributed information retrieval systems: document partitioning and term partitioning. In a document-partitioned system, each of the processors hosts a subset of the documents in the collection, and executes every query against its local sub-collection. In a term-partitioned system, each of the processors hosts a subset of the inverted lists that make up the index of the collection, and serves them to a central machine as they are required for query evaluation. In this paper we introduce a pipelined query-evaluation methodology, based on a term-partitioned index, in which partially evaluated queries are passed amongst the set of processors that host the query terms. This arrangement retains the disk read benefits of term partitioning, but more effectively shares the computational load. We compare the three methodologies experimentally, and show that term distribution is inefficient and scales poorly. The new pipelined approach offers efficient memory utilization and efficient use of disk accesses, but suffers from problems with load balancing between nodes. Until these problems are resolved, document partitioning remains the preferred method. Alistair Moffat was supported by the Australian Research Council, the ARC Special Research Center for Perceptive and Intelligent Machines in Complex Environments, and the NICTA Victoria Laboratory. William Webber and Justin Zobel were supported by the Australian Research Council. Ricardo Baeza-Yates was supported by Grant P01-029-F from Millennium Initiative of Mideplan, Chile; and by the University of Melbourne as a visiting scholar at the time this project was undertaken.     

Impact of question-answering tasks on search processes and reading comprehension

This study examined the effect of (a) high- and low-level questions and (b) reading the text before the questions asked on performance, delayed text recall, and deep text comprehension, as well as on specific text-inspection patterns. Participants were 37 undergraduate students who answered either high- or low-level questions using the software Read&Answer to read and answer questions on the computer screen. Additionally, half of the sample read first a text and then answered the questions (reading-first condition), whereas the other half answered the questions without having read the text in advance (no-reading-first condition). All participants had the text available to search for the answer. Results indicated that high-level questions facilitated deep comprehension but not immediate performance or delayed recall of text, independently of the reading condition, and that high- and low-level questions differentially affected text-inspection patterns.     

A multilevel analysis of late entry in Nigeria

This paper explores how factors operating at the state and community levels are associated with the prevalence of late school enrolment in Nigeria. We investigate the following three research themes. First, whether late entry varies across states and across communities and how much of this variation can be explained by the composition of communities and states. Second, whether community wealth and state wealth are stronger predictors of late entry than household wealth. Third, whether wealth has different effects on the probability of late school entry in different states and in different communities within states. Using the 2010 Nigeria Education Data Survey, and a three-level multilevel logit model, we show that a substantial proportion of the variation in late school entry is explained by community- and state-level factors. We further find that children living in specific communities in Nigeria may face a double disadvantage in terms of on-time school enrolment. In these communities, not only is there a large proportion of children entering school late, but also household wealth has strong effects on the probability that a child is sent to school on on time. These communities could be targeted for educational interventions aimed at promoting on-time school enrolment.     

Advances in urinary protein biomarkers for urogenital and non-urogenital pathologies

The discovery of protein biomarkers that reflect the biological state of the body is of vital importance to disease management. Urine is an ideal source of biomarkers that provides a non-invasive approach to diagnosis, prognosis and prediction of diseases. Consequently, the study of the human urinary proteome has increased dramatically over the last 10 years, with many studies being published. This review focuses on urinary protein biomarkers that have shown potential, in initial studies, for diseases affecting the urogenital tract, specifically chronic kidney disease and prostate cancer, as well as other non-urogenital pathologies such as breast cancer, diabetes, atherosclerosis and osteoarthritis. PubMed was searched for peer-reviewed literature on the subject, published in the last 10 years. The keywords used were “urine, biomarker, protein, and/or prostate cancer/breast cancer/chronic kidney disease/diabetes/atherosclerosis/osteoarthritis”. Original studies on the subject, as well as a small number of reviews, were analysed including the strengths and weaknesses, and we summarized the performance of biomarkers that demonstrated potential. One of the biggest challenges found is that biomarkers are often shared by several pathologies so are not specific to one disease. Therefore, the trend is shifting towards implementing a panel of biomarkers, which may increase specificity. Although there have been many advances in urinary proteomics, these have not resulted in similar advancements in clinical practice due to high costs and the lack of large data sets. In order to translate these potential biomarkers to clinical practice, vigorous validation is needed, with input from industry or large collaborative studies.Key words: urine, protein, biomarker     

The relevance of intellectual capital disclosure: empirical evidence from Spanish universities

The reporting of intellectual capital (IC) in higher education institutions is of vital importance, mainly owing to the fact that knowledge is the main output and input in these institutions. Furthermore, the increasing social concern about establishing procedures of accountability and ensuring information transparency in public universities prompted us to consider the need to disclose information on their IC. In this study, we developed a questionnaire that was sent to members of the Social Councils of Spanish public universities in order to identify the intangible elements about which university stakeholders demand most information. It was also used to see which groups of universities have stakeholders who are more proactive to the disclosure of IC. Our results represent a starting point for public universities to identify the most commonly requested information about IC. Universities can then develop an IC report that fits their own characteristics and environment.     

Kinematic,kinetic and EMG analysis of four front crawl flip turn techniques

This study aimed to analyse the kinematic, kinetic and electromyographic characteristics of four front crawl flip turn technique variants. The variants distinguished from each other by differences in body position (i.e. dorsal, lateral, ventral) during rolling, wall support, pushing and gliding phases. Seventeen highly trained swimmers (17.9 ± 3.2 years old) participated in interventional sessions and performed three trials of each variant, being monitored with a 3-D video system, a force platform and an electromyography (EMG) system. Studied variables: rolling time and distance, wall support time, push-off time, peak force and horizontal impulse at wall support and push-off, centre of mass horizontal velocity at the end of the push-off, gliding time, centre of mass depth, distance, average and final velocity during gliding, total turn time and electrical activity of Gastrocnemius Medialis, Tibialis Anterior, Biceps Femoris and Vastus Lateralis muscles. Depending on the variant, total turn time ranged from 2.37 ± 0.32 to 2.43 ± 0.33 s, push-off force from 1.86 ± 0.33 to 1.92 ± 0.26 BW and centre of mass velocity during gliding from 1.78 ± 0.21 to 1.94 ± 0.22 m · s^?1. The variants were not distinguishable in terms of kinematical, kinetic and EMG parameters during the rolling, wall support, pushing and gliding phases.