Energy-Efficient In-Memory Data Stores on Hybrid Memory Hierarchies

Increasingly large amounts of data are stored in main memory of data center servers. However, DRAM-based memory is an important consumer of energy and is unlikely to scale in the future. Various byte-addressable non-volatile memory (NVM) technologies promise high density and near-zero static energy, however they suffer from increased latency and increased dynamic energy consumption.

This paper proposes to leverage a hybrid memory architecture, consisting of both DRAM and NVM, by novel, application-level data management policies that decide to place data on DRAM vs. NVM. We analyze modern column-oriented and key-value data stores and demonstrate the feasibility of application-level data management. Cycle-accurate simulation confirms that our methodology reduces the energy with least performance degradation as compared to the current state-of-the-art hardware or OS approaches. Moreover, we utilize our techniques to apportion DRAM and NVM memory sizes for these workloads.

Impact of smoking and smoking cessation on cardiovascular events and mortality among older adults: meta-analysis of individual participant data from prospective cohort studies of the CHANCES consortium

OBJECTIVE: To investigate the impact of smoking and smoking cessation on cardiovascular mortality, acute coronary events, and stroke events in people aged 60 and older, and to calculate and report risk advancement periods for cardiovascular mortality in addition to traditional epidemiological relative risk measures.

DESIGN: Individual participant meta-analysis using data from 25 cohorts participating in the CHANCES consortium. Data were harmonised, analysed separately employing Cox proportional hazard regression models, and combined by meta-analysis.

RESULTS: Overall, 503,905 participants aged 60 and older were included in this study, of whom 37,952 died from cardiovascular disease. Random effects meta-analysis of the association of smoking status with cardiovascular mortality yielded a summary hazard ratio of 2.07 (95% CI 1.82 to 2.36) for current smokers and 1.37 (1.25 to 1.49) for former smokers compared with never smokers. Corresponding summary estimates for risk advancement periods were 5.50 years (4.25 to 6.75) for current smokers and 2.16 years (1.38 to 2.39) for former smokers. The excess risk in smokers increased with cigarette consumption in a dose-response manner, and decreased continuously with time since smoking cessation in former smokers. Relative risk estimates for acute coronary events and for stroke events were somewhat lower than for cardiovascular mortality, but patterns were similar.

CONCLUSIONS: Our study corroborates and expands evidence from previous studies in showing that smoking is a strong independent risk factor of cardiovascular events and mortality even at older age, advancing cardiovascular mortality by more than five years, and demonstrating that smoking cessation in these age groups is still beneficial in reducing the excess risk.

Data-Augmentation for Reducing Dataset Bias in Person Re-identification

In this paper we explore ways to address the issue of dataset bias in person re-identification by using data augmentation to increase the variability of the available datasets, and we introduce a novel data augmentation method for re-identification based on changing the image background. We show that use of data augmentation can improve the cross-dataset generalisation of convolutional network based re-identification systems, and that changing the image background yields further improvements.

Iso-Quality of Service: Fairly Ranking Servers for Real-Time Data Analytics

We present a mathematically rigorous Quality-of-Service (QoS) metric which relates the achievable quality of service metric (QoS) for a real-time analytics service to the server energy cost of offering the service. Using a new iso-QoS evaluation methodology, we scale server resources to meet QoS targets and directly rank the servers in terms of their energy-efficiency and by extension cost of ownership. Our metric and method are platform-independent and enable fair comparison of datacenter compute servers with significant architectural diversity, including micro-servers. We deploy our metric and methodology to compare three servers running financial option pricing workloads on real-life market data. We find that server ranking is sensitive to data inputs and desired QoS level and that although scale-out micro-servers can be up to two times more energy-efficient than conventional heavyweight servers for the same target QoS, they are still six times less energy efficient than high-performance computational accelerators.

A Lightweight Tool for Anomaly Detection in Cloud Data Centres

Cloud data centres are critical business infrastructures and the fastest growing service providers. Detecting anomalies in Cloud data centre operation is vital. Given the vast complexity of the data centre system software stack, applications and workloads, anomaly detection is a challenging endeavour. Current tools for detecting anomalies often use machine learning techniques, application instance behaviours or system metrics distribu- tion, which are complex to implement in Cloud computing environments as they require training, access to application-level data and complex processing. This paper presents LADT, a lightweight anomaly detection tool for Cloud data centres that uses rigorous correlation of system metrics, implemented by an efficient corre- lation algorithm without need for training or complex infrastructure set up. LADT is based on the hypothesis that, in an anomaly-free system, metrics from data centre host nodes and virtual machines (VMs) are strongly correlated. An anomaly is detected whenever correlation drops below a threshold value. We demonstrate and evaluate LADT using a Cloud environment, where it shows that the hosting node I/O operations per second (IOPS) are strongly correlated with the aggregated virtual machine IOPS, but this correlation vanishes when an application stresses the disk, indicating a node-level anomaly.

Assessment of Atomic Data: Problems and Solutions

For the reliable analysis and modeling of astrophysical, laser-produced, and fusion plasmas, atomic data are required for a number of parameters, including energy levels, radiative rates, and electron impact excitation rates. Such data are desired for a range of elements (H to W) and their many ions. However, measurements of atomic data, mainly for radiative and excitation rates, are not feasible for many species, and therefore, calculations are needed. For some ions (such as of C, Fe, and Kr), there is a variety of calculations available in the literature, but often, they differ significantly from one another. Therefore, there is a great demand from the user community to have data "assessed" for accuracy so that they can be confidently applied to the modeling of plasmas. In this paper we highlight the difficulties in assessing atomic data and offer some solutions for improving the accuracy of calculated results.

Hui and Walter's latent-class model extended to estimate diagnostic test properties from surveillance data: a latent model for latent data

Diagnostic test sensitivity and specificity are probabilistic estimates with far reaching implications for disease control, management and genetic studies. In the absence of 'gold standard' tests, traditional Bayesian latent class models may be used to assess diagnostic test accuracies through the comparison of two or more tests performed on the same groups of individuals. The aim of this study was to extend such models to estimate diagnostic test parameters and true cohort-specific prevalence, using disease surveillance data. The traditional Hui-Walter latent class methodology was extended to allow for features seen in such data, including (i) unrecorded data (i.e. data for a second test available only on a subset of the sampled population) and (ii) cohort-specific sensitivities and specificities. The model was applied with and without the modelling of conditional dependence between tests. The utility of the extended model was demonstrated through application to bovine tuberculosis surveillance data from Northern and the Republic of Ireland. Simulation coupled with re-sampling techniques, demonstrated that the extended model has good predictive power to estimate the diagnostic parameters and true herd-level prevalence from surveillance data. Our methodology can aid in the interpretation of disease surveillance data, and the results can potentially refine disease control strategies.

Convergent genetic and expression data implicate immunity in Alzheimer's disease

Background: Late-onset Alzheimer's disease (AD) is heritable with 20 genes showing genome-wide association in the International Genomics of Alzheimer's Project (IGAP). To identify the biology underlying the disease, we extended these genetic data in a pathway analysis.

Methods: The ALIGATOR and GSEA algorithms were used in the IGAP data to identify associated functional pathways and correlated gene expression networks in human brain.

Results: ALIGATOR identified an excess of curated biological pathways showing enrichment of association. Enriched areas of biology included the immune response (P = 3.27 X 10(-12) after multiple testing correction for pathways), regulation of endocytosis (P = 1.31 X 10(-11)), cholesterol transport (P = 2.96 X 10(-9)), and proteasome-ubiquitin activity (P = 1.34 X 10(-6)). Correlated gene expression analysis identified four significant network modules, all related to the immune response (corrected P = .002-.05).

Conclusions: The immime response, regulation of endocytosis, cholesterol transport, and protein ubiquitination represent prime targets for AD therapeutics. (C) 2015 Published by Elsevier Inc. on behalf of The Alzheimer's Association.

Robust bimodal person identification using face and speech with limited training data and corruption of both modalities

This paper presents a novel method of audio-visual fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there is a limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new representation and a modified cosine similarity are introduced for combining and comparing bimodal features with limited training data as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal data set created from the SPIDRE and AR databases with variable noise corruption of speech and occlusion in the face images. The new method has demonstrated improved recognition accuracy.

Cinema-going at the Southampton Odeon: Exhibition Data and Popular Taste

Cinemagoing data from the 1970s. A case study of the data from the Southampton Odeon

PESSTO: Survey description and products from the first data release by the Public ESO Spectroscopic Survey of Transient Objects

Context. The Public European Southern Observatory Spectroscopic Survey of Transient Objects (PESSTO) began as a public spectroscopic survey in April 2012. PESSTO classifies transients from publicly available sources and wide-field surveys, and selects science targets for detailed spectroscopic and photometric follow-up. PESSTO runs for nine months of the year, January - April and August - December inclusive, and typically has allocations of 10 nights per month. 

Aims. We describe the data reduction strategy and data products that are publicly available through the ESO archive as the Spectroscopic Survey data release 1 (SSDR1). 

Methods. PESSTO uses the New Technology Telescope with the instruments EFOSC2 and SOFI to provide optical and NIR spectroscopy and imaging. We target supernovae and optical transients brighter than 20.5^m for classification. Science targets are selected for follow-up based on the PESSTO science goal of extending knowledge of the extremes of the supernova population. We use standard EFOSC2 set-ups providing spectra with resolutions of 13-18 Å between 3345-9995 Å. A subset of the brighter science targets are selected for SOFI spectroscopy with the blue and red grisms (0.935-2.53 μm and resolutions 23-33 Å) and imaging with broadband JHK<inf>s</inf> filters. 

Results. This first data release (SSDR1) contains flux calibrated spectra from the first year (April 2012-2013). A total of 221 confirmed supernovae were classified, and we released calibrated optical spectra and classifications publicly within 24 h of the data being taken (via WISeREP). The data in SSDR1 replace those released spectra. They have more reliable and quantifiable flux calibrations, correction for telluric absorption, and are made available in standard ESO Phase 3 formats. We estimate the absolute accuracy of the flux calibrations for EFOSC2 across the whole survey in SSDR1 to be typically ∼15%, although a number of spectra will have less reliable absolute flux calibration because of weather and slit losses. Acquisition images for each spectrum are available which, in principle, can allow the user to refine the absolute flux calibration. The standard NIR reduction process does not produce high accuracy absolute spectrophotometry but synthetic photometry with accompanying JHK<inf>s</inf> imaging can improve this. Whenever possible, reduced SOFI images are provided to allow this. 

Conclusions. Future data releases will focus on improving the automated flux calibration of the data products. The rapid turnaround between discovery and classification and access to reliable pipeline processed data products has allowed early science papers in the first few months of the survey.