Showing papers by "University of Oxford published in 2017"
Theo Vos1, Amanuel Alemu Abajobir, Kalkidan Hassen Abate2, Cristiana Abbafati3 +775 more•Institutions (305)
TL;DR: The Global Burden of Diseases, Injuries, and Risk Factors Study 2016 (GBD 2016) provides a comprehensive assessment of prevalence, incidence, and years lived with disability (YLDs) for 328 causes in 195 countries and territories from 1990 to 2016.
10,401 citations
Posted Content•
TL;DR: Squeeze-and-excitation (SE) as mentioned in this paper adaptively recalibrates channel-wise feature responses by explicitly modeling interdependencies between channels, which can be stacked together to form SENet architectures.
Abstract: The central building block of convolutional neural networks (CNNs) is the convolution operator, which enables networks to construct informative features by fusing both spatial and channel-wise information within local receptive fields at each layer. A broad range of prior research has investigated the spatial component of this relationship, seeking to strengthen the representational power of a CNN by enhancing the quality of spatial encodings throughout its feature hierarchy. In this work, we focus instead on the channel relationship and propose a novel architectural unit, which we term the "Squeeze-and-Excitation" (SE) block, that adaptively recalibrates channel-wise feature responses by explicitly modelling interdependencies between channels. We show that these blocks can be stacked together to form SENet architectures that generalise extremely effectively across different datasets. We further demonstrate that SE blocks bring significant improvements in performance for existing state-of-the-art CNNs at slight additional computational cost. Squeeze-and-Excitation Networks formed the foundation of our ILSVRC 2017 classification submission which won first place and reduced the top-5 error to 2.251%, surpassing the winning entry of 2016 by a relative improvement of ~25%. Models and code are available at this https URL.
5,411 citations
21 Jul 2017
TL;DR: In this article, a Two-Stream Inflated 3D ConvNet (I3D) is proposed to learn seamless spatio-temporal feature extractors from video while leveraging successful ImageNet architecture designs and their parameters.
Abstract: The paucity of videos in current action classification datasets (UCF-101 and HMDB-51) has made it difficult to identify good video architectures, as most methods obtain similar performance on existing small-scale benchmarks. This paper re-evaluates state-of-the-art architectures in light of the new Kinetics Human Action Video dataset. Kinetics has two orders of magnitude more data, with 400 human action classes and over 400 clips per class, and is collected from realistic, challenging YouTube videos. We provide an analysis on how current architectures fare on the task of action classification on this dataset and how much performance improves on the smaller benchmark datasets after pre-training on Kinetics. We also introduce a new Two-Stream Inflated 3D ConvNet (I3D) that is based on 2D ConvNet inflation: filters and pooling kernels of very deep image classification ConvNets are expanded into 3D, making it possible to learn seamless spatio-temporal feature extractors from video while leveraging successful ImageNet architecture designs and even their parameters. We show that, after pre-training on Kinetics, I3D models considerably improve upon the state-of-the-art in action classification, reaching 80.2% on HMDB-51 and 97.9% on UCF-101.
5,073 citations
TL;DR: In this paper, a Gaussian process classifier was used to estimate the probability of computerisation for 702 detailed occupations, and the expected impacts of future computerisation on US labour market outcomes, with the primary objective of analyzing the number of jobs at risk and the relationship between an occupations probability of computing, wages and educational attainment.
4,853 citations
TL;DR: Patients treated with canagliflozin had a lower risk of cardiovascular events than those who received placebo but a greater risk of amputation, primarily at the level of the toe or metatarsal.
Abstract: BackgroundCanagliflozin is a sodium–glucose cotransporter 2 inhibitor that reduces glycemia as well as blood pressure, body weight, and albuminuria in people with diabetes. We report the effects of treatment with canagliflozin on cardiovascular, renal, and safety outcomes. MethodsThe CANVAS Program integrated data from two trials involving a total of 10,142 participants with type 2 diabetes and high cardiovascular risk. Participants in each trial were randomly assigned to receive canagliflozin or placebo and were followed for a mean of 188.2 weeks. The primary outcome was a composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. ResultsThe mean age of the participants was 63.3 years, 35.8% were women, the mean duration of diabetes was 13.5 years, and 65.6% had a history of cardiovascular disease. The rate of the primary outcome was lower with canagliflozin than with placebo (occurring in 26.9 vs. 31.5 participants per 1000 patient-years; hazard ratio, 0.86; 95% c...
4,842 citations
Ashkan Afshin, Mohammad H. Forouzanfar, Marissa B Reitsma, Patrick J Sur +164 more•Institutions (70)
TL;DR: The rapid increase in the prevalence and disease burden of elevated BMI highlights the need for continued focus on surveillance of BMI and identification, implementation, and evaluation of evidence‐based interventions to address this problem.
Abstract: BACKGROUND Although the rising pandemic of obesity has received major attention in many countries, the effects of this attention on trends and the disease burden of obesity remain uncertain. METHOD ...
4,519 citations
University of Udine1, University of Lugano2, École Polytechnique Fédérale de Lausanne3, Leipzig University4, University of Paris5, University of North Texas6, Princeton University7, National Research Council8, International School for Advanced Studies9, Cornell University10, University of Lincoln11, University of Milan12, École Polytechnique13, International Centre for Theoretical Physics14, University of Paderborn15, University of Oxford16, Jožef Stefan Institute17, University of Padua18, Sapienza University of Rome19, Vietnam Academy of Science and Technology20, University of British Columbia21, Centre national de la recherche scientifique22, University of Lorraine23, University of Zurich24, École Normale Supérieure25, Université Paris-Saclay26, Wake Forest University27, Temple University28
TL;DR: Recent extensions and improvements are described, covering new methodologies and property calculators, improved parallelization, code modularization, and extended interoperability both within the distribution and with external software.
Abstract: Quantum ESPRESSO is an integrated suite of open-source computer codes for quantum simulations of materials using state-of-the-art electronic-structure techniques, based on density-functional theory, density-functional perturbation theory, and many-body perturbation theory, within the plane-wave pseudopotential and projector-augmented-wave approaches Quantum ESPRESSO owes its popularity to the wide variety of properties and processes it allows to simulate, to its performance on an increasingly broad array of hardware architectures, and to a community of researchers that rely on its capabilities as a core open-source development platform to implement their ideas In this paper we describe recent extensions and improvements, covering new methodologies and property calculators, improved parallelization, code modularization, and extended interoperability both within the distribution and with external software
3,638 citations
Simon I. Hay, Amanuel Alemu Abajobir1, Kalkidan Hassen Abate2, Cristiana Abbafati3 +800 more•Institutions (32)
TL;DR: At a global level, DALYs and HALE continue to show improvements and the importance of continued health interventions, which has changed in most locations in pace with the gross domestic product per person, education, and family planning.
3,029 citations
Valery L. Feigin1, Amanuel Alemu Abajobir2, Kalkidan Hassen Abate3, Foad Abd-Allah4 +267 more•Institutions (138)
TL;DR: The Global Burden of Diseases, Injuries, and Risk Factors (GBD) study provides such information but does not routinely aggregate results that are of interest to clinicians specialising in neurological conditions as discussed by the authors.
Abstract: Summary Background Comparable data on the global and country-specific burden of neurological disorders and their trends are crucial for health-care planning and resource allocation. The Global Burden of Diseases, Injuries, and Risk Factors (GBD) Study provides such information but does not routinely aggregate results that are of interest to clinicians specialising in neurological conditions. In this systematic analysis, we quantified the global disease burden due to neurological disorders in 2015 and its relationship with country development level. Methods We estimated global and country-specific prevalence, mortality, disability-adjusted life-years (DALYs), years of life lost (YLLs), and years lived with disability (YLDs) for various neurological disorders that in the GBD classification have been previously spread across multiple disease groupings. The more inclusive grouping of neurological disorders included stroke, meningitis, encephalitis, tetanus, Alzheimer's disease and other dementias, Parkinson's disease, epilepsy, multiple sclerosis, motor neuron disease, migraine, tension-type headache, medication overuse headache, brain and nervous system cancers, and a residual category of other neurological disorders. We also analysed results based on the Socio-demographic Index (SDI), a compound measure of income per capita, education, and fertility, to identify patterns associated with development and how countries fare against expected outcomes relative to their level of development. Findings Neurological disorders ranked as the leading cause group of DALYs in 2015 (250·7 [95% uncertainty interval (UI) 229·1 to 274·7] million, comprising 10·2% of global DALYs) and the second-leading cause group of deaths (9·4 [9·1 to 9·7] million], comprising 16·8% of global deaths). The most prevalent neurological disorders were tension-type headache (1505·9 [UI 1337·3 to 1681·6 million cases]), migraine (958·8 [872·1 to 1055·6] million), medication overuse headache (58·5 [50·8 to 67·4 million]), and Alzheimer's disease and other dementias (46·0 [40·2 to 52·7 million]). Between 1990 and 2015, the number of deaths from neurological disorders increased by 36·7%, and the number of DALYs by 7·4%. These increases occurred despite decreases in age-standardised rates of death and DALYs of 26·1% and 29·7%, respectively; stroke and communicable neurological disorders were responsible for most of these decreases. Communicable neurological disorders were the largest cause of DALYs in countries with low SDI. Stroke rates were highest at middle levels of SDI and lowest at the highest SDI. Most of the changes in DALY rates of neurological disorders with development were driven by changes in YLLs. Interpretation Neurological disorders are an important cause of disability and death worldwide. Globally, the burden of neurological disorders has increased substantially over the past 25 years because of expanding population numbers and ageing, despite substantial decreases in mortality rates from stroke and communicable neurological disorders. The number of patients who will need care by clinicians with expertise in neurological conditions will continue to grow in coming decades. Policy makers and health-care providers should be aware of these trends to provide adequate services. Funding Bill & Melinda Gates Foundation.
2,995 citations
University of Udine1, University of Lugano2, École Polytechnique Fédérale de Lausanne3, Leipzig University4, University of Paris5, University of North Texas6, Princeton University7, National Research Council8, International School for Advanced Studies9, Cornell University10, University of Lincoln11, University of Milan12, École Polytechnique13, International Centre for Theoretical Physics14, University of Paderborn15, University of Oxford16, Jožef Stefan Institute17, University of Padua18, Sapienza University of Rome19, Vietnam Academy of Science and Technology20, University of British Columbia21, University of Lorraine22, Centre national de la recherche scientifique23, University of Zurich24, École Normale Supérieure25, Université Paris-Saclay26, Wake Forest University27, Temple University28
TL;DR: Quantum ESPRESSO as discussed by the authors is an integrated suite of open-source computer codes for quantum simulations of materials using state-of-the-art electronic-structure techniques, based on density functional theory, density functional perturbation theory, and many-body perturbations theory, within the plane-wave pseudo-potential and projector-augmented-wave approaches.
Abstract: Quantum ESPRESSO is an integrated suite of open-source computer codes for quantum simulations of materials using state-of-the art electronic-structure techniques, based on density-functional theory, density-functional perturbation theory, and many-body perturbation theory, within the plane-wave pseudo-potential and projector-augmented-wave approaches. Quantum ESPRESSO owes its popularity to the wide variety of properties and processes it allows to simulate, to its performance on an increasingly broad array of hardware architectures, and to a community of researchers that rely on its capabilities as a core open-source development platform to implement theirs ideas. In this paper we describe recent extensions and improvements, covering new methodologies and property calculators, improved parallelization, code modularization, and extended interoperability both within the distribution and with external software.
2,818 citations
Icahn School of Medicine at Mount Sinai1, Pure Earth2, World Bank3, University of Arizona4, McGill University5, Indian Ministry of Environment and Forests6, Qatar Airways7, University of Health Sciences Antigua8, Ludwig Maximilian University of Munich9, Johns Hopkins University10, Boston College11, Chulabhorn Research Institute12, University of Maryland, College Park13, University of Ghana14, Centro Nacional de Investigaciones Cardiovasculares15, University of Chicago16, University of London17, University of Oxford18, Indian Institute of Technology Delhi19, Simon Fraser University20, Consortium of Universities for Global Health21, University of Ottawa22, Columbia University23, Stockholm Resilience Centre24, Massachusetts Institute of Technology25, University of Queensland26, University of California, Berkeley27, New York University28, National Institutes of Health29, Public Health Research Institute30, United Nations Industrial Development Organization31, Renmin University of China32
TL;DR: This book is dedicated to the memory of those who have served in the armed forces and their families during the conflicts of the twentieth century.
Gregory A. Roth1, Catherine O. Johnson1, Amanuel Alemu Abajobir2, Foad Abd-Allah3 +170 more•Institutions (99)
TL;DR: The GBD (Global Burden of Disease) 2015 study integrated data on disease incidence, prevalence, and mortality to produce consistent, up-to-date estimates for cardiovascular burden, finding that CVDs remain a major cause of health loss for all regions of the world.
01 Jul 2017
TL;DR: In this paper, the authors introduce a new training strategy, iCaRL, that allows learning in such a class-incremental way: only the training data for a small number of classes has to be present at the same time and new classes can be added progressively.
Abstract: A major open problem on the road to artificial intelligence is the development of incrementally learning systems that learn about more and more concepts over time from a stream of data. In this work, we introduce a new training strategy, iCaRL, that allows learning in such a class-incremental way: only the training data for a small number of classes has to be present at the same time and new classes can be added progressively. iCaRL learns strong classifiers and a data representation simultaneously. This distinguishes it from earlier works that were fundamentally limited to fixed data representations and therefore incompatible with deep learning architectures. We show by experiments on CIFAR-100 and ImageNet ILSVRC 2012 data that iCaRL can learn many classes incrementally over a long period of time where other strategies quickly fail.
01 Jan 2017
TL;DR: Greenwood, Greenwood, Christine Oliver, Kerstin Sahlin and Roy Suddaby as mentioned in this paper discuss the work of Meanings in Institutional Processes and Thinking, and discuss the importance of meaning in organizational processes and thinking.
Abstract: Introduction - Royston Greenwood, Christine Oliver, Kerstin Sahlin and Roy Suddaby PART ONE: FOUNDATIONAL THEMES Legitimacy in Organizational Institutionalism - David L Deephouse and Marc Suchman Isomorphism, Diffusion and Decoupling - Eva Boxenbaum and Stefan Jonsson Institutional Logics - Patricia H Thornton and William Ocasio Organizational Fields - Melissa Wooten and Andrew J Hoffman Past, Present and Future PART TWO: INSTITUTIONAL DYNAMICS The Work of Meanings in Institutional Processes and Thinking - Tammar B Zilber Power, Institutions and Organizations - Thomas B Lawrence Institutional Entrepreneurship - Cynthia Hardy and Steve Maguire Circulating Ideas - Kerstin Sahlin and Linda Wedlin Imitation, Translation and Editing Organizational Implications of Institutional Pluralism - Matthew S Kraatz and Emily S. Block Microfoundations of Institutional Theory - Walter W Powell and Jeannette A. Colyvas Institutions and Transnationalization - Marie-Laure Djelic and Sigrid Quack PART THREE: APPLICATIONS Traditions as Institutionalized Practice - M Tina Dacin and Peter A Dacin Implications for de-institutionalization New Forms as Settlements - Hayagreeva Rao and Martin Kenney Social Movements and Failed Institutionalization - Gerald F Davis and Peter J J Anderson Corporate, (Non)Response to the AIDS Epidemic Institutions and Corporate Governance - Peer C Fiss PART FOUR: INTERFACES Beyond Constraint - Mary Ann Glynn How Institutions Enable Identities Institutionalism and the Professions - Kevin T Leicht and Mary L Fennell Institutionalism and Globalization Studies - Gili S Drori Organizational Institutionalism and Sociology - C R Hinings and Pamela S Tolbert A Reflection Coalface Institutionalism - Stephen R Barley New Sociology of Knowledge - Renate E Meyer Historical Legacy and Contributions Systems Theory, Societal Contexts and Organizational Heterogeneity - Raimund Hasse and Georg Kr cken Charting Progress at the Nexus of Institutional Theory and Economics - Peter W Roberts Ecologists and Institutionalists - Heather A Haveman and Robert J David Friends or Foes? Networks and Institutions - Jason Owen-Smith and Walter W Powell Institutional-Level Learning - Pamela Haunschild and David Chandler Learning as a Source of Institutional Change Social Movements and Institutional Analysis - Marc Schneiberg and Michael Lounsbury Examining 'Institutionalization' - David J Cooper, Mahmoud Ezzamel and Hugh Willmott A Critical Theoretic Perspective Taking Social Construction Seriously - Nelson Phillips and Namrata Malhotra Extending the Discursive Approach in Institutional Theory Institutional Leadership - Marvin Washington, Kimberly B Boal and John N Davis Past, Present and Future PART FIVE: REFLECTIONS Is the New Institutionalism a Theory? - Donald Palmer, Nicole Biggart and Brian Dick How to Misuse Institutions and Get Away with It - Barbara Czarniawska Some Reflections on Institutional Theory(ies) Been There, Done That, Moving on - Paul Hirsch Reflections on Institutional Theory's Continuing Evolution Reflections on Institutional Theories of Organizations - John W Meyer
Posted Content•
TL;DR: Relation Network (RN) as mentioned in this paper learns to learn a deep distance metric to compare a small number of images within episodes, each of which is designed to simulate the few-shot setting.
Abstract: We present a conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each. Our method, called the Relation Network (RN), is trained end-to-end from scratch. During meta-learning, it learns to learn a deep distance metric to compare a small number of images within episodes, each of which is designed to simulate the few-shot setting. Once trained, a RN is able to classify images of new classes by computing relation scores between query images and the few examples of each new class without further updating the network. Besides providing improved performance on few-shot learning, our framework is easily extended to zero-shot learning. Extensive experiments on five benchmarks demonstrate that our simple approach provides a unified and effective approach for both of these two tasks.
TL;DR: This work argues for the adoption of measures to optimize key elements of the scientific process: methods, reporting and dissemination, reproducibility, evaluation and incentives, in the hope that this will facilitate action toward improving the transparency, reproducible and efficiency of scientific research.
Abstract: Improving the reliability and efficiency of scientific research will increase the credibility of the published scientific literature and accelerate discovery. Here we argue for the adoption of measures to optimize key elements of the scientific process: methods, reporting and dissemination, reproducibility, evaluation and incentives. There is some evidence from both simulations and empirical studies supporting the likely effectiveness of these measures, but their broad adoption by researchers, institutions, funders and journals will require iterative evaluation and improvement. We discuss the goals of these measures, and how they can be implemented, in the hope that this will facilitate action toward improving the transparency, reproducibility and efficiency of scientific research.
Posted Content•
TL;DR: I3D models considerably improve upon the state-of-the-art in action classification, reaching 80.2% on HMDB-51 and 97.9% on UCF-101 after pre-training on Kinetics, and a new Two-Stream Inflated 3D Conv net that is based on 2D ConvNet inflation is introduced.
Abstract: The paucity of videos in current action classification datasets (UCF-101 and HMDB-51) has made it difficult to identify good video architectures, as most methods obtain similar performance on existing small-scale benchmarks. This paper re-evaluates state-of-the-art architectures in light of the new Kinetics Human Action Video dataset. Kinetics has two orders of magnitude more data, with 400 human action classes and over 400 clips per class, and is collected from realistic, challenging YouTube videos. We provide an analysis on how current architectures fare on the task of action classification on this dataset and how much performance improves on the smaller benchmark datasets after pre-training on Kinetics.
We also introduce a new Two-Stream Inflated 3D ConvNet (I3D) that is based on 2D ConvNet inflation: filters and pooling kernels of very deep image classification ConvNets are expanded into 3D, making it possible to learn seamless spatio-temporal feature extractors from video while leveraging successful ImageNet architecture designs and even their parameters. We show that, after pre-training on Kinetics, I3D models considerably improve upon the state-of-the-art in action classification, reaching 80.9% on HMDB-51 and 98.0% on UCF-101.
TL;DR: Deep reinforcement learning (DRL) is poised to revolutionize the field of artificial intelligence (AI) and represents a step toward building autonomous systems with a higher-level understanding of the visual world as discussed by the authors.
Abstract: Deep reinforcement learning (DRL) is poised to revolutionize the field of artificial intelligence (AI) and represents a step toward building autonomous systems with a higherlevel understanding of the visual world. Currently, deep learning is enabling reinforcement learning (RL) to scale to problems that were previously intractable, such as learning to play video games directly from pixels. DRL algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of RL, then progress to the main streams of value-based and policy-based methods. Our survey will cover central algorithms in deep RL, including the deep Q-network (DQN), trust region policy optimization (TRPO), and asynchronous advantage actor critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via RL. To conclude, we describe several current areas of research within the field.
University College London1, London Research Institute2, University of Leicester3, Francis Crick Institute4, University of Manchester5, University of Birmingham6, Glenfield Hospital7, Middlesex University8, Princess Alexandra Hospital NHS Trust9, Cardiff University10, University of Oxford11, Max Delbrück Center for Molecular Medicine12, Katholieke Universiteit Leuven13
TL;DR: Intratumor heterogeneity mediated through chromosome instability was associated with an increased risk of recurrence or death, a finding that supports the potential value of chromosome instability as a prognostic predictor.
Abstract: BackgroundAmong patients with non–small-cell lung cancer (NSCLC), data on intratumor heterogeneity and cancer genome evolution have been limited to small retrospective cohorts. We wanted to prospectively investigate intratumor heterogeneity in relation to clinical outcome and to determine the clonal nature of driver events and evolutionary processes in early-stage NSCLC. MethodsIn this prospective cohort study, we performed multiregion whole-exome sequencing on 100 early-stage NSCLC tumors that had been resected before systemic therapy. We sequenced and analyzed 327 tumor regions to define evolutionary histories, obtain a census of clonal and subclonal events, and assess the relationship between intratumor heterogeneity and recurrence-free survival. ResultsWe observed widespread intratumor heterogeneity for both somatic copy-number alterations and mutations. Driver mutations in EGFR, MET, BRAF, and TP53 were almost always clonal. However, heterogeneous driver alterations that occurred later in evolution w...
TL;DR: The number of adults with raised blood pressure increased from 594 million in 1975 to 1·13 billion in 2015, with the increase largely in low-income and middle-income countries, and the contributions of changes in prevalence versus population growth and ageing to the increase.
New York University1, National and Kapodistrian University of Athens2, University of Barcelona3, Texas Oncology4, Complutense University of Madrid5, Aix-Marseille University6, University of Toronto7, University of Oxford8, University of Queensland9, University of Colorado Denver10, Emory University11, Cross Cancer Institute12, Centre Hospitalier Universitaire de Toulouse13, Georgetown University14, University of Sydney15, University of Washington16, University of Paris17, Nagoya University18, Seoul National University19, Bristol-Myers Squibb20, The Royal Marsden NHS Foundation Trust21
TL;DR: Among patients undergoing resection of stage IIIB, IIIC, or IV melanoma, adjuvant therapy with nivolumab resulted in significantly longer recurrence‐free survival and a lower rate of grade 3 or 4 adverse events than adjuant therapy with ipilimumab.
Abstract: BackgroundNivolumab and ipilimumab are immune checkpoint inhibitors that have been approved for the treatment of advanced melanoma. In the United States, ipilimumab has also been approved as adjuvant therapy for melanoma on the basis of recurrence-free and overall survival rates that were higher than those with placebo in a phase 3 trial. We wanted to determine the efficacy of nivolumab versus ipilimumab for adjuvant therapy in patients with resected advanced melanoma. MethodsIn this randomized, double-blind, phase 3 trial, we randomly assigned 906 patients (≥15 years of age) who were undergoing complete resection of stage IIIB, IIIC, or IV melanoma to receive an intravenous infusion of either nivolumab at a dose of 3 mg per kilogram of body weight every 2 weeks (453 patients) or ipilimumab at a dose of 10 mg per kilogram every 3 weeks for four doses and then every 12 weeks (453 patients). The patients were treated for a period of up to 1 year or until disease recurrence, a report of unacceptable toxic ef...
TL;DR: It is argued that while this law will pose large challenges for industry, it highlights opportunities for computer scientists to take the lead in designing algorithms and evaluation frameworks which avoid discrimination and enable explanation.
Abstract: We summarize the potential impact that the European Union’s new General Data Protection Regulation will have on the routine use of machine learning algorithms. Slated to take effect as law across the EU in 2018, it will restrict automated individual decision-making (that is, algorithms that make decisions based on user-level predictors) which “significantly affect” users. The law will also effectively create a “right to explanation,” whereby a user can ask for an explanation of an algorithmic decision that was made about them. We argue that while this law will pose large challenges for industry, it highlights opportunities for computer scientists to take the lead in designing algorithms and evaluation frameworks which avoid discrimination and enable explanation.
Institute for Health Metrics and Evaluation1, Jimma University2, University of Alabama at Birmingham3, University of Oxford4, University of Cartagena5, University College London6, Harvard University7, Wellcome Trust8, University of Canterbury9, Madawalabu University10, Ottawa Hospital Research Institute11, University of Valencia12, Auckland University of Technology13, University of Coimbra14, Bielefeld University15, Mekelle University16, University of Massachusetts Boston17, University of Western Australia18, Harry Perkins Institute of Medical Research19, Heidelberg University20, New Generation University College21, Southern University College22, Simmons College23, Brown University24, University of Melbourne25, University of São Paulo26, University of Adelaide27, National Institutes of Health28, Columbia University29, Southern Illinois University Carbondale30, Ministry of Health and Social Welfare31, Teikyo University32, University of British Columbia33, Marshall University34, South African Medical Research Council35, Addis Ababa University36, Arba Minch University37, Northumbria University38, University of Edinburgh39, James Cook University40, Monash University41, University of Calgary42, University of Copenhagen43, University of Warwick44, National Research University – Higher School of Economics45, Duke University46, Northwestern University47
TL;DR: In international surveys, although there is uncertainty in some estimates, the rate of elevatedSBP (≥110-115 and ≥140 mm Hg) increased substantially between 1990 and 2015, and DALYs and deaths associated with elevated SBP also increased.
Abstract: Importance Elevated systolic blood (SBP) pressure is a leading global health risk. Quantifying the levels of SBP is important to guide prevention policies and interventions. Objective To estimate the association between SBP of at least 110 to 115 mm Hg and SBP of 140 mm Hg or higher and the burden of different causes of death and disability by age and sex for 195 countries and territories, 1990-2015. Design A comparative risk assessment of health loss related to SBP. Estimated distribution of SBP was based on 844 studies from 154 countries (published 1980-2015) of 8.69 million participants. Spatiotemporal Gaussian process regression was used to generate estimates of mean SBP and adjusted variance for each age, sex, country, and year. Diseases with sufficient evidence for a causal relationship with high SBP (eg, ischemic heart disease, ischemic stroke, and hemorrhagic stroke) were included in the primary analysis. Main Outcomes and Measures Mean SBP level, cause-specific deaths, and health burden related to SBP (≥110-115 mm Hg and also ≥140 mm Hg) by age, sex, country, and year. Results Between 1990-2015, the rate of SBP of at least 110 to 115 mm Hg increased from 73 119 (95% uncertainty interval [UI], 67 949-78 241) to 81 373 (95% UI, 76 814-85 770) per 100 000, and SBP of 140 mm Hg or higher increased from 17 307 (95% UI, 17 117-17 492) to 20 526 (95% UI, 20 283-20 746) per 100 000. The estimated annual death rate per 100 000 associated with SBP of at least 110 to 115 mm Hg increased from 135.6 (95% UI, 122.4-148.1) to 145.2 (95% UI 130.3-159.9) and the rate for SBP of 140 mm Hg or higher increased from 97.9 (95% UI, 87.5-108.1) to 106.3 (95% UI, 94.6-118.1). Loss of disability-adjusted life-years (DALYs) associated with SBP of at least 110 to 115 mm Hg increased from 148 million (95% UI, 134-162 million) to 211 million (95% UI, 193-231 million), and for SBP of 140 mm Hg or higher, the loss increased from 95.9 million (95% UI, 87.0-104.9 million) to 143.0 million (95% UI, 130.2-157.0 million). The largest numbers of SBP-related deaths were caused by ischemic heart disease (4.9 million [95% UI, 4.0-5.7 million]; 54.5%), hemorrhagic stroke (2.0 million [95% UI, 1.6-2.3 million]; 58.3%), and ischemic stroke (1.5 million [95% UI, 1.2-1.8 million]; 50.0%). In 2015, China, India, Russia, Indonesia, and the United States accounted for more than half of the global DALYs related to SBP of at least 110 to 115 mm Hg. Conclusions and Relevance In international surveys, although there is uncertainty in some estimates, the rate of elevated SBP (≥110-115 and ≥140 mm Hg) increased substantially between 1990 and 2015, and DALYs and deaths associated with elevated SBP also increased. Projections based on this sample suggest that in 2015, an estimated 3.5 billion adults had SBP of at least 110 to 115 mm Hg and 874 million adults had SBP of 140 mm Hg or higher.
Anglia Ruskin University1, University of Oxford2, Heidelberg University3, L V Prasad Eye Institute4, Massachusetts Eye and Ear Infirmary5, Nova Southeastern University6, University of KwaZulu-Natal7, Brien Holden Vision Institute8, Flinders University9, University of New South Wales10, Royal Liverpool University Hospital11, World Health Organization12, National University of Singapore13, University of Melbourne14, Selçuk University15, University of Burgundy16, University of Miami17, University of Adelaide18, Queen's University Belfast19, Harvard University20, The George Institute for Global Health21, University of Washington22, University of Michigan23, Universiti Tunku Abdul Rahman24, University of Alabama25, National Institutes of Health26, Johns Hopkins University27, University of São Paulo28, Henry Ford Health System29, University College London30, University of Nairobi31, University of Georgia32, University of Utah33, Federal University of São Paulo34, Yale University35, Alberta Children's Hospital36, University of Pennsylvania37, Medical College of Wisconsin38, Novartis39, University of Udine40, University of Illinois at Urbana–Champaign41, Royal Children's Hospital42, University of Missouri43, University of Milan44, Centers for Disease Control and Prevention45, Singapore National Eye Center46, Icahn School of Medicine at Mount Sinai47, Mayo Clinic48, Pan American Health Organization49, University of Indonesia50, University of Crete51, Erasmus University Rotterdam52, University of Southern California53, University of Florence54, Stellenbosch University55, Capital Medical University56, Leipzig University57, Moorfields Eye Hospital58
TL;DR: There is an ongoing reduction in the age-standardised prevalence of blindness and visual impairment, yet the growth and ageing of the world's population is causing a substantial increase in number of people affected, highlighting the need to scale up vision impairment alleviation efforts at all levels.
TL;DR: The addition of abiraterone acetate and prednisone to androgen‐deprivation therapy significantly increased overall survival and radiographic progression‐free survival in men with newly diagnosed, metastatic, castration‐sensitive prostate cancer.
Abstract: BackgroundAbiraterone acetate, a drug that blocks endogenous androgen synthesis, plus prednisone is indicated for metastatic castration-resistant prostate cancer We evaluated the clinical benefit of abiraterone acetate plus prednisone with androgen-deprivation therapy in patients with newly diagnosed, metastatic, castration-sensitive prostate cancer MethodsIn this double-blind, placebo-controlled, phase 3 trial, we randomly assigned 1199 patients to receive either androgen-deprivation therapy plus abiraterone acetate (1000 mg daily, given once daily as four 250-mg tablets) plus prednisone (5 mg daily) (the abiraterone group) or androgen-deprivation therapy plus dual placebos (the placebo group) The two primary end points were overall survival and radiographic progression-free survival ResultsAfter a median follow-up of 304 months at a planned interim analysis (after 406 patients had died), the median overall survival was significantly longer in the abiraterone group than in the placebo group (not rea
[...]
Massachusetts Institute of Technology1, Broad Institute2, Howard Hughes Medical Institute3, University of Cambridge4, European Bioinformatics Institute5, Wellcome Trust Sanger Institute6, Harvard University7, Weizmann Institute of Science8, University of Zurich9, Laboratory of Molecular Biology10, Utrecht University11, École Polytechnique Fédérale de Lausanne12, University of Pennsylvania13, Heidelberg University14, German Cancer Research Center15, Ludwig Maximilian University of Munich16, John Radcliffe Hospital17, Newcastle University18, Stanford University19, University of Oxford20, University of California, San Francisco21, Allen Institute for Brain Science22, Karolinska Institutet23, Royal Institute of Technology24, Icahn School of Medicine at Mount Sinai25, University of Cape Town26, University Medical Center Groningen27, Radboud University Nijmegen28, Kettering University29, University of Edinburgh30, Babraham Institute31, New York University32, Netherlands Cancer Institute33, Ragon Institute of MGH, MIT and Harvard34, University of Texas Health Science Center at Houston35, Technische Universität München36, Technical University of Denmark37, University of California, Berkeley38, King's College London39, California Institute of Technology40
TL;DR: An open comprehensive reference map of the molecular state of cells in healthy human tissues would propel the systematic study of physiological states, developmental trajectories, regulatory circuitry and interactions of cells, and also provide a framework for understanding cellular dysregulation in human disease.
Abstract: The recent advent of methods for high-throughput single-cell molecular profiling has catalyzed a growing sense in the scientific community that the time is ripe to complete the 150-year-old effort to identify all cell types in the human body. The Human Cell Atlas Project is an international collaborative effort that aims to define all human cell types in terms of distinctive molecular profiles (such as gene expression profiles) and to connect this information with classical cellular descriptions (such as location and morphology). An open comprehensive reference map of the molecular state of cells in healthy human tissues would propel the systematic study of physiological states, developmental trajectories, regulatory circuitry and interactions of cells, and also provide a framework for understanding cellular dysregulation in human disease. Here we describe the idea, its potential utility, early proofs-of-concept, and some design considerations for the Human Cell Atlas, including a commitment to open data, code, and community.
TL;DR: The InTBIR Participants and Investigators have provided informed consent for the study to take place in Poland.
Abstract: Additional co-authors: Endre Czeiter, Marek Czosnyka, Ramon Diaz-Arrastia, Jens P Dreier, Ann-Christine Duhaime, Ari Ercole, Thomas A van Essen, Valery L Feigin, Guoyi Gao, Joseph Giacino, Laura E Gonzalez-Lara, Russell L Gruen, Deepak Gupta, Jed A Hartings, Sean Hill, Ji-yao Jiang, Naomi Ketharanathan, Erwin J O Kompanje, Linda Lanyon, Steven Laureys, Fiona Lecky, Harvey Levin, Hester F Lingsma, Marc Maegele, Marek Majdan, Geoffrey Manley, Jill Marsteller, Luciana Mascia, Charles McFadyen, Stefania Mondello, Virginia Newcombe, Aarno Palotie, Paul M Parizel, Wilco Peul, James Piercy, Suzanne Polinder, Louis Puybasset, Todd E Rasmussen, Rolf Rossaint, Peter Smielewski, Jeannette Soderberg, Simon J Stanworth, Murray B Stein, Nicole von Steinbuchel, William Stewart, Ewout W Steyerberg, Nino Stocchetti, Anneliese Synnot, Braden Te Ao, Olli Tenovuo, Alice Theadom, Dick Tibboel, Walter Videtta, Kevin K W Wang, W Huw Williams, Kristine Yaffe for the InTBIR Participants and Investigators
TL;DR: The membrane raft hypothesis formalized a physicochemical principle for a subtype of lateral membrane heterogeneity, in which the preferential associations between cholesterol and saturated lipids drive the formation of relatively packed membrane domains that selectively recruit certain lipids and proteins.
Abstract: Cellular plasma membranes are laterally heterogeneous, featuring a variety of distinct subcompartments that differ in their biophysical properties and composition. A large number of studies have focused on understanding the basis for this heterogeneity and its physiological relevance. The membrane raft hypothesis formalized a physicochemical principle for a subtype of such lateral membrane heterogeneity, in which the preferential associations between cholesterol and saturated lipids drive the formation of relatively packed (or ordered) membrane domains that selectively recruit certain lipids and proteins. Recent studies have yielded new insights into this mechanism and its relevance in vivo, owing primarily to the development of improved biochemical and biophysical technologies.
21 Jul 2017
TL;DR: In this paper, the Correlation Filter learner is interpreted as a differentiable layer in a deep neural network, which enables learning deep features that are tightly coupled to the correlation filter.
Abstract: The Correlation Filter is an algorithm that trains a linear template to discriminate between images and their translations. It is well suited to object tracking because its formulation in the Fourier domain provides a fast solution, enabling the detector to be re-trained once per frame. Previous works that use the Correlation Filter, however, have adopted features that were either manually designed or trained for a different task. This work is the first to overcome this limitation by interpreting the Correlation Filter learner, which has a closed-form solution, as a differentiable layer in a deep neural network. This enables learning deep features that are tightly coupled to the Correlation Filter. Experiments illustrate that our method has the important practical benefit of allowing lightweight architectures to achieve state-of-the-art performance at high framerates.
TL;DR: Among patients with type 2 diabetes with or without previous cardiovascular disease, the incidence of major adverse cardiovascular events did not differ significantly between patients who received exenatide and those who received placebo.
Abstract: BackgroundThe cardiovascular effects of adding once-weekly treatment with exenatide to usual care in patients with type 2 diabetes are unknown. MethodsWe randomly assigned patients with type 2 diabetes, with or without previous cardiovascular disease, to receive subcutaneous injections of extended-release exenatide at a dose of 2 mg or matching placebo once weekly. The primary composite outcome was the first occurrence of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. The coprimary hypotheses were that exenatide, administered once weekly, would be noninferior to placebo with respect to safety and superior to placebo with respect to efficacy. ResultsIn all, 14,752 patients (of whom 10,782 [73.1%] had previous cardiovascular disease) were followed for a median of 3.2 years (interquartile range, 2.2 to 4.4). A primary composite outcome event occurred in 839 of 7356 patients (11.4%; 3.7 events per 100 person-years) in the exenatide group and in 905 of 7396 patients (12.2...