scispace - formally typeset
Search or ask a question

Showing papers by "National Research University – Higher School of Economics published in 2018"


Journal ArticleDOI
Gregory A. Roth1, Gregory A. Roth2, Degu Abate3, Kalkidan Hassen Abate4  +1025 moreInstitutions (333)
TL;DR: Non-communicable diseases comprised the greatest fraction of deaths, contributing to 73·4% (95% uncertainty interval [UI] 72·5–74·1) of total deaths in 2017, while communicable, maternal, neonatal, and nutritional causes accounted for 18·6% (17·9–19·6), and injuries 8·0% (7·7–8·2).

5,211 citations


Journal ArticleDOI
TL;DR: In this paper, the authors assess the burden of 29 cancer groups over time to provide a framework for policy discussion, resource allocation, and research focus, and evaluate cancer incidence, mortality, years lived with disability, years of life lost, and disability-adjusted life-years (DALYs) for 195 countries and territories by age and sex using the Global Burden of Disease study estimation methods.
Abstract: Importance The increasing burden due to cancer and other noncommunicable diseases poses a threat to human development, which has resulted in global political commitments reflected in the Sustainable Development Goals as well as the World Health Organization (WHO) Global Action Plan on Non-Communicable Diseases. To determine if these commitments have resulted in improved cancer control, quantitative assessments of the cancer burden are required. Objective To assess the burden for 29 cancer groups over time to provide a framework for policy discussion, resource allocation, and research focus. Evidence Review Cancer incidence, mortality, years lived with disability, years of life lost, and disability-adjusted life-years (DALYs) were evaluated for 195 countries and territories by age and sex using the Global Burden of Disease study estimation methods. Levels and trends were analyzed over time, as well as by the Sociodemographic Index (SDI). Changes in incident cases were categorized by changes due to epidemiological vs demographic transition. Findings In 2016, there were 17.2 million cancer cases worldwide and 8.9 million deaths. Cancer cases increased by 28% between 2006 and 2016. The smallest increase was seen in high SDI countries. Globally, population aging contributed 17%; population growth, 12%; and changes in age-specific rates, −1% to this change. The most common incident cancer globally for men was prostate cancer (1.4 million cases). The leading cause of cancer deaths and DALYs was tracheal, bronchus, and lung cancer (1.2 million deaths and 25.4 million DALYs). For women, the most common incident cancer and the leading cause of cancer deaths and DALYs was breast cancer (1.7 million incident cases, 535 000 deaths, and 14.9 million DALYs). In 2016, cancer caused 213.2 million DALYs globally for both sexes combined. Between 2006 and 2016, the average annual age-standardized incidence rates for all cancers combined increased in 130 of 195 countries or territories, and the average annual age-standardized death rates decreased within that timeframe in 143 of 195 countries or territories. Conclusions and Relevance Large disparities exist between countries in cancer incidence, deaths, and associated disability. Scaling up cancer prevention and ensuring universal access to cancer care are required for health equity and to fulfill the global commitments for noncommunicable disease and cancer control.

4,621 citations


Journal ArticleDOI
Jeffrey D. Stanaway1, Ashkan Afshin1, Emmanuela Gakidou1, Stephen S Lim1  +1050 moreInstitutions (346)
TL;DR: This study estimated levels and trends in exposure, attributable deaths, and attributable disability-adjusted life-years (DALYs) by age group, sex, year, and location for 84 behavioural, environmental and occupational, and metabolic risks or groups of risks from 1990 to 2017 and explored the relationship between development and risk exposure.

2,910 citations



Proceedings Article
14 Mar 2018
TL;DR: Stochastic Weight Averaging (SWA) as discussed by the authors is a simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conventional training.
Abstract: Deep neural networks are typically trained by optimizing a loss function with an SGD variant, in conjunction with a decaying learning rate, until convergence. We show that simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conventional training. We also show that this Stochastic Weight Averaging (SWA) procedure finds much flatter solutions than SGD, and approximates the recent Fast Geometric Ensembling (FGE) approach with a single model. Using SWA we achieve notable improvement in test accuracy over conventional SGD training on a range of state-of-the-art residual networks, PyramidNets, DenseNets, and Shake-Shake networks on CIFAR-10, CIFAR-100, and ImageNet. In short, SWA is extremely easy to implement, improves generalization, and has almost no computational overhead.

330 citations


Journal ArticleDOI
TL;DR: This work estimated population in 195 locations by single year of age and single calendar year from 1950 to 2017 with standardised and replicable methods and used the cohort-component method of population projection, with inputs of fertility, mortality, population, and migration data.

287 citations


Journal ArticleDOI
TL;DR: The program to apply modern quantum field theory methods to calculate observables in classical general relativity through a truncation to classical terms of the multigraviton, two-body, on-shell scattering amplitudes between massive fields is outlined.
Abstract: Weoutline the program to apply modern quantum field theory methods to calculate observables in classical general relativity through a truncation to classical terms of the multigraviton, two-body, on-shell scattering amplitudes between massive fields. Since only long-distance interactions corresponding to nonanalytic pieces need to be included, unitarity cuts provide substantial simplifications for both post-Newtonian and post-Minkowskian expansions. We illustrate this quantum field theoretic approach to classical general relativity by computing the interaction potentials to second order in the post-Newtonian expansion, as well as the scattering functions for two massive objects to second order in the post-Minkowskian expansion. We also derive an all-order exact result for gravitational light-by-light scattering.

286 citations


Journal ArticleDOI
Gregory A. Roth1, Gregory A. Roth2, Catherine O. Johnson2, Kalkidan Hassen Abate3, Foad Abd-Allah4, Muktar Beshir Ahmed3, Khurshid Alam5, Tahiya Alam2, Nelson Alvis-Guzman6, Hossein Ansari, Johan Ärnlöv7, Tesfay Mehari Atey8, Ashish Awasthi9, Tadesse Awoke10, Aleksandra Barac11, Till Bärnighausen12, Neeraj Bedi13, Derrick A Bennett14, Isabela M. Benseñor15, Sibhatu Biadgilign, Carlos A Castañeda-Orjuela, Ferrán Catalá-López16, Kairat Davletov17, Samath D Dharmaratne18, Eric L. Ding12, Manisha Dubey19, Emerito Jose A. Faraon20, Talha Farid21, Maryam S. Farvid12, Valery L. Feigin22, João C. Fernandes23, Joseph Frostad2, Alemseged Aregay Gebru8, Johanna M. Geleijnse24, Philimon Gona25, Max Griswold2, Gessessew Bugssa Hailu8, Graeme J. Hankey5, Hamid Yimam Hassen26, Rasmus Havmoeller7, Simon I. Hay2, Susan R. Heckbert2, Caleb Mackay Salpeter Irvine2, Spencer L. James2, Dube Jara27, Amir Kasaeian28, Abdur Rahman Khan21, Sahil Khera29, Abdullah T Khoja30, Jagdish Khubchandani31, Daniel Kim32, Dhaval Kolte33, Dharmesh Kumar Lal9, Anders Larsson34, Shai Linn35, Paulo A. Lotufo15, Hassan Magdy Abd El Razek36, Mohsen Mazidi37, Toni Meier38, Walter Mendoza39, George A. Mensah40, Atte Meretoja41, Haftay Berhane Mezgebe8, Erkin M. Mirrakhimov42, Shafiu Mohammed43, Andrew E. Moran44, Grant Nguyen2, Minh Nguyen2, Kanyin Liane Ong2, Mayowa O. Owolabi45, Martin A Pletcher2, Farshad Pourmalek46, Caroline A. Purcell2, Mostafa Qorbani, Mahfuzar Rahman47, Rajesh Kumar Rai, Usha Ram19, Marissa B Reitsma2, Andre M. N. Renzaho48, Maria Jesus Rios-Blancas, Saeid Safiri49, Joshua A. Salomon12, Benn Sartorius50, Sadaf G. Sepanlou28, Masood Ali Shaikh, Diego Augusto Santos Silva51, Saverio Stranges52, Rafael Tabarés-Seisdedos16, Niguse Tadele Atnafu53, Jarnail Singh Thakur54, Roman Topor-Madry55, Thomas Truelsen56, E. Murat Tuzcu57, Stefanos Tyrovolas58, Kingsley N. Ukwaja, Tommi Vasankari, Vasiliy Victorovich Vlassov59, Stein Emil Vollset60, Tolassa Wakayo3, Robert G. Weintraub61, Charles D.A. Wolfe62, Abdulhalik Workicho3, Gelin Xu63, Simon Yadgir2, Yuichiro Yano64, Paul S. F. Yip65, Naohiro Yonemoto66, Mustafa Z. Younis67, Chuanhua Yu68, Zoubida Zaidi, Maysaa El Sayed Zaki36, Ben Zipkin2, Ashkan Afshin2, Emmanuela Gakidou2, Stephen S Lim2, Ali H. Mokdad2, Mohsen Naghavi2, Theo Vos2, Christopher J L Murray2 
University of Washington1, Institute for Health Metrics and Evaluation2, Jimma University3, Cairo University4, University of Western Australia5, University of Cartagena6, Karolinska Institutet7, Mekelle University8, Public Health Foundation of India9, University of Gondar10, University of Belgrade11, Harvard University12, Jazan University13, University of Oxford14, University of São Paulo15, University of Valencia16, Kazakh National Medical University17, University of Peradeniya18, International Institute for Population Sciences19, University of the Philippines Manila20, University of Louisville21, Auckland University of Technology22, Catholic University of Portugal23, Wageningen University and Research Centre24, University of Massachusetts Boston25, Mizan–Tepi University26, Debre markos University27, Tehran University of Medical Sciences28, New York Medical College29, Islamic University30, Ball State University31, Northeastern University32, Brown University33, Uppsala University34, University of Haifa35, Mansoura University36, Chinese Academy of Sciences37, Martin Luther University of Halle-Wittenberg38, United Nations Population Fund39, National Institutes of Health40, University of Melbourne41, Kyrgyz State Medical Academy42, Ahmadu Bello University43, Columbia University44, University of Ibadan45, University of British Columbia46, BRAC47, University of Sydney48, University of Maragheh49, University of KwaZulu-Natal50, Universidade Federal de Santa Catarina51, University of Western Ontario52, Addis Ababa University53, Post Graduate Institute of Medical Education and Research54, Jagiellonian University Medical College55, University of Copenhagen56, Cleveland Clinic57, Hospital Sant Joan de Déu Barcelona58, National Research University – Higher School of Economics59, Norwegian Institute of Public Health60, Royal Children's Hospital61, King's College London62, Nanjing University63, University of Mississippi Medical Center64, University of Hong Kong65, Kyoto University66, Jackson State University67, Wuhan University68
TL;DR: Large disparities in total burden of CVD persist between US states despite marked improvements in CVD burden, and increases in risk-deleted CVD DALY rates between 2006 and 2016 in 16 states suggest additional unmeasured risks beyond these traditional factors.
Abstract: Importance Cardiovascular disease (CVD) is the leading cause of death in the United States, but regional variation within the United States is large. Comparable and consistent state-level measures of total CVD burden and risk factors have not been produced previously. Objective To quantify and describe levels and trends of lost health due to CVD within the United States from 1990 to 2016 as well as risk factors driving these changes. Design, Setting, and Participants Using the Global Burden of Disease methodology, cardiovascular disease mortality, nonfatal health outcomes, and associated risk factors were analyzed by age group, sex, and year from 1990 to 2016 for all residents in the United States using standardized approaches for data processing and statistical modeling. Burden of disease was estimated for 10 groupings of CVD, and comparative risk analysis was performed. Data were analyzed from August 2016 to July 2017. Exposures Residing in the United States. Main Outcomes and Measures Cardiovascular disease disability-adjusted life-years (DALYs). Results Between 1990 and 2016, age-standardized CVD DALYs for all states decreased. Several states had large rises in their relative rank ordering for total CVD DALYs among states, including Arkansas, Oklahoma, Alabama, Kentucky, Missouri, Indiana, Kansas, Alaska, and Iowa. The rate of decline varied widely across states, and CVD burden increased for a small number of states in the most recent years. Cardiovascular disease DALYs remained twice as large among men compared with women. Ischemic heart disease was the leading cause of CVD DALYs in all states, but the second most common varied by state. Trends were driven by 12 groups of risk factors, with the largest attributable CVD burden due to dietary risk exposures followed by high systolic blood pressure, high body mass index, high total cholesterol level, high fasting plasma glucose level, tobacco smoking, and low levels of physical activity. Increases in risk-deleted CVD DALY rates between 2006 and 2016 in 16 states suggest additional unmeasured risks beyond these traditional factors. Conclusions and Relevance Large disparities in total burden of CVD persist between US states despite marked improvements in CVD burden. Differences in CVD burden are largely attributable to modifiable risk exposures.

261 citations


Proceedings Article
27 Feb 2018
TL;DR: Fast Geometric Ensembling (FGE) as mentioned in this paper uses a simple curve to connect the optima of the loss functions of deep neural networks, over which training and test accuracy are nearly constant.
Abstract: The loss functions of deep neural networks are complex and their geometric properties are not well understood. We show that the optima of these complex loss functions are in fact connected by simple curves, over which training and test accuracy are nearly constant. We introduce a training procedure to discover these high-accuracy pathways between modes. Inspired by this new geometric insight, we also propose a new ensembling method entitled Fast Geometric Ensembling (FGE). Using FGE we can train high-performing ensembles in the time required to train a single model. We achieve improved performance compared to the recent state-of-the-art Snapshot Ensembles, on CIFAR-10, CIFAR-100, and ImageNet.

259 citations


Journal ArticleDOI
TL;DR: In this paper, the authors explore and profile the nature and dynamics of the Quadruple/Quintuple Helix Innovation System Model or Framework (government, university, industry, civil society, environment) as an enabler and enactor of regional co-opetitive entrepreneurial ecosystems which they conceptualize as fractal, multi-level, multimodal, mult-nodal and multi-lateral configurations of dynamic tangible and intangible assets within the resource-based view and the new theory of the growth of the firm.
Abstract: Regions are increasingly being viewed as eco-systemic agglomerations of organizational and institutional entities or stakeholders with socio-technical, socio-economic, and socio-political conflicting as well as converging (co-opetitive) goals, priorities, expectations, and behaviors that they pursue via entrepreneurial development, exploration, exploitation, and deployment actions, reactions and interactions. In this context, our paper aims to explore and profile the nature and dynamics of the Quadruple/Quintuple Helix Innovation System Model or Framework (government, university, industry, civil society, environment) as an enabler and enactor of regional co-opetitive entrepreneurial ecosystems which we conceptualize as fractal, multi-level, multi-modal, multi-nodal, and multi-lateral configurations of dynamic tangible and intangible assets within the resource-based view and the new theory of the growth of the firm. Co-opetitive fractal innovation and entrepreneurship ecosystems are defined and discussed, and examples of regional innovation policies and programs are presented. Furthermore, the concept of multi-level innovation systems is analyzed, taking into account the existence of knowledge clusters and innovation networks, while alternative aggregations of multi-level innovation systems are proposed based on their spatial (geographical) and non-spatial (research-based) functional properties.

253 citations


Posted Content
TL;DR: A benchmarking platform called Molecular Sets (MOSES) is introduced to standardize training and comparison of molecular generative models and suggest to use the results as reference points for further advancements in generative chemistry research.
Abstract: Generative models are becoming a tool of choice for exploring the molecular space. These models learn on a large training dataset and produce novel molecular structures with similar properties. Generated structures can be utilized for virtual screening or training semi-supervised predictive models in the downstream tasks. While there are plenty of generative models, it is unclear how to compare and rank them. In this work, we introduce a benchmarking platform called Molecular Sets (MOSES) to standardize training and comparison of molecular generative models. MOSES provides a training and testing datasets, and a set of metrics to evaluate the quality and diversity of generated structures. We have implemented and compared several molecular generation models and suggest to use our results as reference points for further advancements in generative chemistry research. The platform and source code are available at this https URL.

Posted Content
TL;DR: It is shown that the optima of these complex loss functions are in fact connected by simple curves over which training and test accuracy are nearly constant, and a training procedure is introduced to discover these high-accuracy pathways between modes.
Abstract: The loss functions of deep neural networks are complex and their geometric properties are not well understood. We show that the optima of these complex loss functions are in fact connected by simple curves over which training and test accuracy are nearly constant. We introduce a training procedure to discover these high-accuracy pathways between modes. Inspired by this new geometric insight, we also propose a new ensembling method entitled Fast Geometric Ensembling (FGE). Using FGE we can train high-performing ensembles in the time required to train a single model. We achieve improved performance compared to the recent state-of-the-art Snapshot Ensembles, on CIFAR-10, CIFAR-100, and ImageNet.

Journal ArticleDOI
TL;DR: Female students are less overconfident than male students, their forecasts are more rational, and they are also faster learners in the sense that they adjust their expectations more rapidly.
Abstract: Confidence and overconfidence are essential aspects of human nature, but measuring (over)confidence is not easy. Our approach is to consider students' forecasts of their exam grades. Part of a student's grade expectation is based on the student's previous academic achievements; what remains can be interpreted as (over)confidence. Our results are based on a sample of about 500 second-year undergraduate students enrolled in a statistics course in Moscow. The course contains three exams and each student produces a forecast for each of the three exams. Our models allow us to estimate overconfidence quantitatively. Using these models we find that students' expectations are not rational and that most students are overconfident, in agreement with the general literature. Less obvious is that overconfidence helps: given the same academic achievement students with larger confidence obtain higher exam grades. Female students are less overconfident than male students, their forecasts are more rational, and they are also faster learners in the sense that they adjust their expectations more rapidly.

Journal ArticleDOI
28 Aug 2018-JAMA
TL;DR: Between 195 000 and 276 000 firearm injury deaths globally in 2016 were estimated, the majority of which were firearm homicides, and there was variation among countries and across demographic subgroups.
Abstract: Importance Understanding global variation in firearm mortality rates could guide prevention policies and interventions. Objective To estimate mortality due to firearm injury deaths from 1990 to 2016 in 195 countries and territories. Design, Setting, and Participants This study used deidentified aggregated data including 13 812 location-years of vital registration data to generate estimates of levels and rates of death by age-sex-year-location. The proportion of suicides in which a firearm was the lethal means was combined with an estimate of per capita gun ownership in a revised proxy measure used to evaluate the relationship between availability or access to firearms and firearm injury deaths. Exposures Firearm ownership and access. Main Outcomes and Measures Cause-specific deaths by age, sex, location, and year. Results Worldwide, it was estimated that 251 000 (95% uncertainty interval [UI], 195 000-276 000) people died from firearm injuries in 2016, with 6 countries (Brazil, United States, Mexico, Colombia, Venezuela, and Guatemala) accounting for 50.5% (95% UI, 42.2%-54.8%) of those deaths. In 1990, there were an estimated 209 000 (95% UI, 172 000 to 235 000) deaths from firearm injuries. Globally, the majority of firearm injury deaths in 2016 were homicides (64.0% [95% UI, 54.2%-68.0%]; absolute value, 161 000 deaths [95% UI, 107 000-182 000]); additionally, 27% were firearm suicide deaths (67 500 [95% UI, 55 400-84 100]) and 9% were unintentional firearm deaths (23 000 [95% UI, 18 200-24 800]). From 1990 to 2016, there was no significant decrease in the estimated global age-standardized firearm homicide rate (−0.2% [95% UI, −0.8% to 0.2%]). Firearm suicide rates decreased globally at an annualized rate of 1.6% (95% UI, 1.1-2.0), but in 124 of 195 countries and territories included in this study, these levels were either constant or significant increases were estimated. There was an annualized decrease of 0.9% (95% UI, 0.5%-1.3%) in the global rate of age-standardized firearm deaths from 1990 to 2016. Aggregate firearm injury deaths in 2016 were highest among persons aged 20 to 24 years (for men, an estimated 34 700 deaths [95% UI, 24 900-39 700] and for women, an estimated 3580 deaths [95% UI, 2810-4210]). Estimates of the number of firearms by country were associated with higher rates of firearm suicide ( P R2 = 0.21) and homicide ( P R2 = 0.35). Conclusions and Relevance This study estimated between 195 000 and 276 000 firearm injury deaths globally in 2016, the majority of which were firearm homicides. Despite an overall decrease in rates of firearm injury death since 1990, there was variation among countries and across demographic subgroups.

Journal ArticleDOI
TL;DR: A new generative architecture is proposed, entangled conditional adversarial autoencoder, that generates molecular structures based on various properties, such as activity against a specific protein, solubility, or ease of synthesis, that is applied to generate a novel inhibitor of Janus kinase 3.
Abstract: Modern computational approaches and machine learning techniques accelerate the invention of new drugs. Generative models can discover novel molecular structures within hours, while conventional drug discovery pipelines require months of work. In this article, we propose a new generative architecture, entangled conditional adversarial autoencoder, that generates molecular structures based on various properties, such as activity against a specific protein, solubility, or ease of synthesis. We apply the proposed model to generate a novel inhibitor of Janus kinase 3, implicated in rheumatoid arthritis, psoriasis, and vitiligo. The discovered molecule was tested in vitro and showed good activity and selectivity.

Journal ArticleDOI
TL;DR: DEREPLICATOR+ is reported, an algorithm that improves on the previous approaches for identifying peptidic natural products, and extends them for identification of polyketides, terpenes, benzenoids, alkaloids, flavonoids, and other classes of natural products.
Abstract: Natural products have traditionally been rich sources for drug discovery. In order to clear the road toward the discovery of unknown natural products, biologists need dereplication strategies that identify known ones. Here we report DEREPLICATOR+, an algorithm that improves on the previous approaches for identifying peptidic natural products, and extends them for identification of polyketides, terpenes, benzenoids, alkaloids, flavonoids, and other classes of natural products. We show that DEREPLICATOR+ can search all spectra in the recently launched Global Natural Products Social molecular network and identify an order of magnitude more natural products than previous dereplication efforts. We further demonstrate that DEREPLICATOR+ enables cross-validation of genome-mining and peptidogenomics/glycogenomics results.

Proceedings Article
01 May 2018
TL;DR: The authors proposed an alternative approach to computing reparameterization gradients based on implicit differentiation and demonstrate its broader applicability by applying it to Gamma, Beta, Dirichlet, and von Mises distributions.
Abstract: By providing a simple and efficient way of computing low-variance gradients of continuous random variables, the reparameterization trick has become the technique of choice for training a variety of latent variable models. However, it is not applicable to a number of important continuous distributions. We introduce an alternative approach to computing reparameterization gradients based on implicit differentiation and demonstrate its broader applicability by applying it to Gamma, Beta, Dirichlet, and von Mises distributions, which cannot be used with the classic reparameterization trick. Our experiments show that the proposed approach is faster and more accurate than the existing gradient estimators for these distributions.

Journal ArticleDOI
TL;DR: If current trends in tuberculosis incidence continue, few countries are likely to meet the SDG target to end the tuberculosis epidemic by 2030, and several regions had higher rates of age-standardised incidence and mortality than expected on the basis of their SDI levels in 2016.
Abstract: Summary Background Although a preventable and treatable disease, tuberculosis causes more than a million deaths each year. As countries work towards achieving the Sustainable Development Goal (SDG) target to end the tuberculosis epidemic by 2030, robust assessments of the levels and trends of the burden of tuberculosis are crucial to inform policy and programme decision making. We assessed the levels and trends in the fatal and non-fatal burden of tuberculosis by drug resistance and HIV status for 195 countries and territories from 1990 to 2016. Methods We analysed 15 943 site-years of vital registration data, 1710 site-years of verbal autopsy data, 764 site-years of sample-based vital registration data, and 361 site-years of mortality surveillance data to estimate mortality due to tuberculosis using the Cause of Death Ensemble model. We analysed all available data sources, including annual case notifications, prevalence surveys, population-based tuberculin surveys, and estimated tuberculosis cause-specific mortality to generate internally consistent estimates of incidence, prevalence, and mortality using DisMod-MR 2.1, a Bayesian meta-regression tool. We assessed how the burden of tuberculosis differed from the burden predicted by the Socio-demographic Index (SDI), a composite indicator of income per capita, average years of schooling, and total fertility rate. Findings Globally in 2016, among HIV-negative individuals, the number of incident cases of tuberculosis was 9·02 million (95% uncertainty interval [UI] 8·05–10·16) and the number of tuberculosis deaths was 1·21 million (1·16–1·27). Among HIV-positive individuals, the number of incident cases was 1·40 million (1·01–1·89) and the number of tuberculosis deaths was 0·24 million (0·16–0·31). Globally, among HIV-negative individuals the age-standardised incidence of tuberculosis decreased annually at a slower rate (–1·3% [–1·5 to −1·2]) than mortality did (–4·5% [–5·0 to −4·1]) from 2006 to 2016. Among HIV-positive individuals during the same period, the rate of change in annualised age-standardised incidence was −4·0% (–4·5 to −3·7) and mortality was −8·9% (–9·5 to −8·4). Several regions had higher rates of age-standardised incidence and mortality than expected on the basis of their SDI levels in 2016. For drug-susceptible tuberculosis, the highest observed-to-expected ratios were in southern sub-Saharan Africa (13·7 for incidence and 14·9 for mortality), and the lowest ratios were in high-income North America (0·4 for incidence) and Oceania (0·3 for mortality). For multidrug-resistant tuberculosis, eastern Europe had the highest observed-to-expected ratios (67·3 for incidence and 73·0 for mortality), and high-income North America had the lowest ratios (0·4 for incidence and 0·5 for mortality). Interpretation If current trends in tuberculosis incidence continue, few countries are likely to meet the SDG target to end the tuberculosis epidemic by 2030. Progress needs to be accelerated by improving the quality of and access to tuberculosis diagnosis and care, by developing new tools, scaling up interventions to prevent risk factors for tuberculosis, and integrating control programmes for tuberculosis and HIV. Funding Bill & Melinda Gates Foundation.

Journal ArticleDOI
TL;DR: A database of historical and archaeological information from 30 regions around the world over the last 10,000 years revealed that characteristics, such as social scale, economy, features of governance, and information systems, show strong evolutionary relationships with each other and that complexity of a society across different world regions can be meaningfully measured using a single principal component of variation.
Abstract: Do human societies from around the world exhibit similarities in the way that they are structured, and show commonalities in the ways that they have evolved? These are long-standing questions that have proven difficult to answer. To test between competing hypotheses, we constructed a massive repository of historical and archaeological information known as “Seshat: Global History Databank.” We systematically coded data on 414 societies from 30 regions around the world spanning the last 10,000 years. We were able to capture information on 51 variables reflecting nine characteristics of human societies, such as social scale, economy, features of governance, and information systems. Our analyses revealed that these different characteristics show strong relationships with each other and that a single principal component captures around three-quarters of the observed variation. Furthermore, we found that different characteristics of social complexity are highly predictable across different world regions. These results suggest that key aspects of social organization are functionally related and do indeed coevolve in predictable ways. Our findings highlight the power of the sciences and humanities working together to rigorously test hypotheses about general rules that may have shaped human history.

Journal ArticleDOI
TL;DR: In this paper, an extremely bright gamma-ray burst (GRB) 160625B was observed in both gamma and optical wavelengths, with three isolated episodes separated by long quiescent intervals, with the durations of each sub-burst being approximately 0.8, 35, and 212 seconds.
Abstract: The ejecta composition is an open question in gamma-ray burst (GRB) physics 1 . Some GRBs possess a quasi-thermal spectral component in the time-resolved spectral analysis 2 , suggesting a hot fireball origin. Others show a featureless non-thermal spectrum known as the Band function3–5, consistent with a synchrotron radiation origin5,6 and suggesting that the jet is Poynting-flux dominated at the central engine and probably in the emission region as well7,8. There are also bursts showing a sub-dominant thermal component and a dominant synchrotron component 9 , suggesting a probable hybrid jet composition 10 . Here, we report an extraordinarily bright GRB 160625B, simultaneously observed in gamma-ray and optical wavelengths, whose prompt emission consists of three isolated episodes separated by long quiescent intervals, with the durations of each sub-burst being approximately 0.8 s, 35 s and 212 s, respectively. Its high brightness (with isotropic peak luminosity Lp,iso ≈ 4 × 1053 erg s−1) allows us to conduct detailed time-resolved spectral analysis in each episode, from precursor to main burst and to extended emission. The spectral properties of the first two sub-bursts are distinctly different, allowing us to observe the transition from thermal to non-thermal radiation between well-separated emission episodes within a single GRB. Such a transition is a clear indication of the change of jet composition from a fireball to a Poynting-flux-dominated jet. The extremely bright GRB 160625B, consisting of three sub-bursts separated by quiescent intervals, shows a transition from thermal to non-thermal radiation that indicates a change of jet composition from a fireball to a Poynting-flux-dominated jet.

Journal ArticleDOI
Roel Aaij, Bernardo Adeva1, Marco Adinolfi2, C. A. Aidala3  +858 moreInstitutions (72)
TL;DR: The doubly charmed baryon decay Ξcc++→Ξc+π+ is observed for the first time, with a statistical significance of 5.9σ as mentioned in this paper.
Abstract: The doubly charmed baryon decay Ξcc++→Ξc+π+ is observed for the first time, with a statistical significance of 5.9σ, confirming a recent observation of the baryon in the Λc+K-π+π+ final state. The data sample used corresponds to an integrated luminosity of 1.7 fb-1, collected by the LHCb experiment in pp collisions at a center-of-mass energy of 13 TeV. The Ξcc++ mass is measured to be 3620.6±1.5(stat)±0.4(syst)±0.3(Ξc+) MeV/c2 and is consistent with the previous result. The ratio of branching fractions between the decay modes is measured to be [B(Ξcc++→Ξc+π+)×B(Ξc+→pK-π+)]/[B(Ξcc++→Λc+K-π+π+)×B(Λc+→pK-π+)]=0.035±0.009(stat)±0.003(syst).

Journal ArticleDOI
TL;DR: In this article, the existence of four independent gradings in colored HOMFLYPT homology, and qualitative predictions of various interesting structures and symmetries in the colored homology of arbitrary knots, are discussed.
Abstract: We conjecture the existence of four independent gradings in colored HOMFLYPT homology, and make qualitative predictions of various interesting structures and symmetries in the colored homology of arbitrary knots. We propose an explicit conjectural description for the rectangular colored homology of torus knots, and identify the new gradings in this context. While some of these structures have a natural interpretation in the physical realization of knot homologies based on counting supersymmetric configurations (BPS states, instantons, and vortices), others are completely new. They suggest new geometric and physical realizations of colored HOMFLYPT homology as the Hochschild homology of the category of branes in a Landau–Ginzburg B-model or, equivalently, in the mirror A-model. Supergroups and supermanifolds are surprisingly ubiquitous in all aspects of this work.

Proceedings Article
03 Jul 2018
TL;DR: In this article, two algorithms for approximating the general optimal transport (OT) distance between two discrete distributions of size n, up to accuracy δ(n 2 ) were presented.
Abstract: We analyze two algorithms for approximating the general optimal transport (OT) distance between two discrete distributions of size $n$, up to accuracy $\varepsilon$. For the first algorithm, which is based on the celebrated Sinkhorn's algorithm, we prove the complexity bound $\widetilde{O}\left({n^2/\varepsilon^2}\right)$ arithmetic operations. For the second one, which is based on our novel Adaptive Primal-Dual Accelerated Gradient Descent (APDAGD) algorithm, we prove the complexity bound $\widetilde{O}\left(\min\left\{n^{9/4}/\varepsilon, n^{2}/\varepsilon^2 \right\}\right)$ arithmetic operations. Both bounds have better dependence on $\varepsilon$ than the state-of-the-art result given by $\widetilde{O}\left({n^2/\varepsilon^3}\right)$. Our second algorithm not only has better dependence on $\varepsilon$ in the complexity bound, but also is not specific to entropic regularization and can solve the OT problem with different regularizers.

Posted Content
TL;DR: It is shown that simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conventional training, and Stochastic Weight Averaging (SWA) is extremely easy to implement, improves generalization, and has almost no computational overhead.
Abstract: Deep neural networks are typically trained by optimizing a loss function with an SGD variant, in conjunction with a decaying learning rate, until convergence. We show that simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conventional training. We also show that this Stochastic Weight Averaging (SWA) procedure finds much flatter solutions than SGD, and approximates the recent Fast Geometric Ensembling (FGE) approach with a single model. Using SWA we achieve notable improvement in test accuracy over conventional SGD training on a range of state-of-the-art residual networks, PyramidNets, DenseNets, and Shake-Shake networks on CIFAR-10, CIFAR-100, and ImageNet. In short, SWA is extremely easy to implement, improves generalization, and has almost no computational overhead.

Journal ArticleDOI
TL;DR: An empirically derived dimensional model of inhibition characterizing neural systems underlying different aspects of inhibitory mechanisms is provided, offering a fundamental framework to advance current understanding of inhibition and provides new insights for future clinical research into disorders with different types of inhibition‐related dysfunctions.
Abstract: Inhibitory control is the stopping of a mental process with or without intention, conceptualized as mental suppression of competing information because of limited cognitive capacity. Inhibitory control dysfunction is a core characteristic of many major psychiatric disorders. Inhibition is generally thought to involve the prefrontal cortex; however, a single inhibitory mechanism is insufficient for interpreting the heterogeneous nature of human cognition. It remains unclear whether different dimensions of inhibitory processes-specifically cognitive inhibition, response inhibition, and emotional interference-rely on dissociated neural systems. We conducted systematic meta-analyses of fMRI studies in the BrainMap database supplemented by PubMed using whole-brain activation likelihood estimation. A total of 66 study experiments including 1,447 participants and 987 foci revealed that while the left anterior insula was concordant in all inhibitory dimensions, cognitive inhibition reliably activated specific dorsal frontal inhibitory system, engaging dorsal anterior cingulate, dorsolateral prefrontal cortex, and parietal areas, whereas emotional interference reliably implicated a ventral inhibitory system, involving the ventral surface of the inferior frontal gyrus and the amygdala. Response inhibition showed concordant clusters in the fronto-striatal system, including the dorsal anterior cingulate region and extended supplementary motor areas, the dorsal and ventral lateral prefrontal cortex, basal ganglia, midbrain regions, and parietal regions. We provide an empirically derived dimensional model of inhibition characterizing neural systems underlying different aspects of inhibitory mechanisms. This study offers a fundamental framework to advance current understanding of inhibition and provides new insights for future clinical research into disorders with different types of inhibition-related dysfunctions.

Journal ArticleDOI
TL;DR: Results of preliminary studies on how neural networks can be utilized to path planning on square grids show that the agent using neural Q-learning algorithm robustly learns to achieve the goal on small maps and demonstrate promising results on the maps have ben never seen by him before.

Journal ArticleDOI
TL;DR: The study suggests that the power conversion efficiencies of solar cells can be controlled by changing the halide composition in all-inorganic perovskite QDs.
Abstract: We demonstrate that halide content strongly affects nonradiative electron–hole recombination in all-inorganic perovskite quantum dots (QDs). Using time domain density functional theory and nonadiabatic molecular dynamics, we show that replacing half of the bromines with iodines in a CsPbBr3 QD extends the charge carrier lifetime by a factor of 5, while complete replacement extends the lifetime by a factor of 8. Doping with iodines decreases the nonadiabatic charge–phonon coupling because iodines are heavier and slower than bromines and because the overlap between the electron and hole wave functions is reduced. In general, the nonradiative electron–hole recombination proceeds slowly, on a nanosecond time scale, due to small sub-1 meV nonadiabatic coupling and short sub-10 fs coherence times. The obtained recombination times and their dependence on the halogen content show excellent agreement with experiments. Our study suggests that the power conversion efficiencies of solar cells can be controlled by cha...

Journal ArticleDOI
TL;DR: In this article, the existence of non-reciprocal waveguide electromagnetic modes in ferromagnetic Weyl semimetal films in the Voigt configuration is predicted, thanks to the strong dielectric response caused by the gapless Weyl spectrum and the large Berry curvature.
Abstract: The propagation of light in Weyl semimetal films is analyzed. The magnetic family of these materials is known by anomalous Hall effect, which, being enhanced by the large Berry curvature, allows one to create strong gyrotropic and nonreciprocity effects without external magnetic field. The existence of nonreciprocal waveguide electromagnetic modes in ferromagnetic Weyl semimetal films in the Voigt configuration is predicted. Thanks to the strong dielectric response caused by the gapless Weyl spectrum and the large Berry curvature, ferromagnetic Weyl semimetals combine the best waveguide properties of magnetic dielectrics or semiconductors with strong anomalous Hall effect in ferromagnets. The magnitude of the nonreciprocity depends both on the internal Weyl semimetal properties, the separation of Weyl nodes, and the external factor, the optical contrast between the media surrounding the film. By tuning the Fermi level in Weyl semimetals, one can vary the operation frequencies of the waveguide modes in THz and mid-IR ranges. Our findings pave the way to the design of compact, tunable, and effective nonreciprocal optical elements.


Journal ArticleDOI
N. Agafonova1, R. Brugnera2, N. Di Marco, D. A. Podgrudkov3, L. Patrizii, Naotaka Naganawa4, H. Shibuya5, Andrey Ustyuzhanin6, T. V. Shchedrina1, A. Buonaura, I. Kreslo7, P. Del Amo Sanchez8, F. Fornari9, A. Bertolin, Seok Kim10, A. Sadovsky11, A. Schembri, Shigeki Aoki12, N. M. Okateva1, E. Voevodina, E. Shibayama12, N. G. Polukhina1, N. G. Polukhina13, F. Mizutani12, Timur Dzhatdoev3, Tomoko Ariga7, Tomoko Ariga14, L. Votano, Hiroki Rokujo4, M. Pozzato9, T. Matsuo5, M. T. Muciaccia15, C. S. Yoon10, A. Hollnagel16, S. Dusini, Mario Stipčević, C. Jollet17, K. Ozaki12, B. Klicek, G. Mandrioli, Alessandro Paoloni, L. Consiglio, L. Paparella15, S. Ogawa5, F. Pupilli, D. Duchesneau8, Kunihiro Morishima4, A. Garfagnini2, L. Pasqualini9, Antonio Ereditato7, Anselmo Meregaglia17, Osamu Sato4, Mitsuhiro Nakamura4, G. Grella18, A. M. Guler19, Simona Maria Stellacci18, Pierre Vilain20, S. Tufanli7, A. Di Crescenzo, U. Kose, Koichi Kodama21, J. Ebert16, P. Strolin, M. Roda2, M. Kamiscioglu19, B. D. Park10, Caren Hagner16, Chiara Sirignano2, A. Alexandrov, M. Tenti, G. De Lellis, C. Kamiscioglu19, C. Kamiscioglu22, Motoaki Miyanishi4, Satoru Takahashi12, N. Mauri9, T. Hayakawa4, P. Monacelli, I. Shakiryanova1, A. Iuliano, S. Dmitrievsky11, Tatsuhiro Naka4, S. Buontempo, N. I. Starkov1, S. Vasina11, Nicola D'Ambrosio, F. Laudisio2, M. Spinetti, Giuliana Galati, F. Terranova23, G. Rosa, Alessandro Pastore, Ante Ljubičić, J. Favier8, T. M. Roganova3, R. A. Fini, A. Sotnikov11, Björn Wonsak16, Adele Lauria, Toshiyuki Nakano4, Krešimir Jakovčić, C. Gustavino, S. Gorbunov1, H. Pessard8, T. Fukuda4, M. Malenica, O. G. Ryazhskaya1, Cristiano Bozza18, Katsumi Ishiguro4, S. Mikado24, Artem Chukanov11, C. Pistillo7, A. Longhin2, D. Di Ferdinando, Maria Cristina Montesi, Valeri Tioukov, S. Simone15, E. Medinaceli2, Y. A. Gornushkin11, J. Goldberg25, J. L. Vuilleumier7, V. Gentile, M. Dracos17, V. A. Matveev1, K. Niwa4, Nobuko Kitagawa4, Masahiro Komatsu4, G. Sirri, Luca Stanco, P. F. Loverre, T. Hara12, M. De Serio15, M. Chernyavskiy1, Gaston Wilquet20, A. M. Anokhina3, T. Shiraishi4, A. S. Malgin1, A. Olchevsky11, Akitaka Ariga7 
TL;DR: The OPERA experiment was designed to study νμ}→ντ} oscillations in the appearance mode in the CERN to Gran Sasso Neutrino beam (CNGS) as mentioned in this paper.
Abstract: The OPERA experiment was designed to study ν_{μ}→ν_{τ} oscillations in the appearance mode in the CERN to Gran Sasso Neutrino beam (CNGS). In this Letter, we report the final analysis of the full data sample collected between 2008 and 2012, corresponding to 17.97×10^{19} protons on target. Selection criteria looser than in previous analyses have produced ten ν_{τ} candidate events, thus reducing the statistical uncertainty in the measurement of the oscillation parameters and of ν_{τ} properties. A multivariate approach for event identification has been applied to the candidate events and the discovery of ν_{τ} appearance is confirmed with an improved significance level of 6.1σ. |Δm_{32}^{2}| has been measured, in appearance mode, with an accuracy of 20%. The measurement of the ν_{τ} charged-current cross section, for the first time with a negligible contamination from ν[over ¯]_{τ}, and the first direct evidence for the ν_{τ} lepton number are also reported.