Griffith University

Wang J., Sun K., Cheng T., Jiang B., Deng C., Zhao Y., Liu D., Mu Y., Tan M., Wang X., Liu W., Xiao B.

High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation through a subnetwork that is formed by connecting high-to-low resolution convolutions in series (e.g., ResNet, VGGNet), and then recover the high-resolution representation from the encoded low-resolution representation. Instead, our proposed network, named as High-Resolution Network (HRNet), maintains high-resolution representations through the whole process. There are two key characteristics: (i) Connect the high-to-low resolution convolution streams in parallel and (ii) repeatedly exchange the information across resolutions. The benefit is that the resulting representation is semantically richer and spatially more precise. We show the superiority of the proposed HRNet in a wide range of applications, including human pose estimation, semantic segmentation, and object detection, suggesting that the HRNet is a stronger backbone for computer vision problems. All the codes are available at https://github.com/HRNet .

Li Z., Liu F., Yang W., Peng S., Zhou J.

A convolutional neural network (CNN) is one of the most significant networks in the deep learning field. Since CNN made impressive achievements in many areas, including but not limited to computer vision and natural language processing, it attracted much attention from both industry and academia in the past few years. The existing reviews mainly focus on CNN’s applications in different scenarios without considering CNN from a general perspective, and some novel ideas proposed recently are not covered. In this review, we aim to provide some novel ideas and prospects in this fast-growing field. Besides, not only 2-D convolution but also 1-D and multidimensional ones are involved. First, this review introduces the history of CNN. Second, we provide an overview of various convolutions. Third, some classic and advanced CNN models are introduced; especially those key points making them reach state-of-the-art results. Fourth, through experimental analysis, we draw some conclusions and provide several rules of thumb for functions and hyperparameter selection. Fifth, the applications of 1-D, 2-D, and multidimensional convolution are covered. Finally, some open issues and promising directions for CNN are discussed as guidelines for future work.

Tudi M., Daniel Ruan H., Wang L., Lyu J., Sadler R., Connell D., Chu C., Phung D.T.

Pesticides are indispensable in agricultural production. They have been used by farmers to control weeds and insects, and their remarkable increases in agricultural products have been reported. The increase in the world’s population in the 20th century could not have been possible without a parallel increase in food production. About one-third of agricultural products are produced depending on the application of pesticides. Without the use of pesticides, there would be a 78% loss of fruit production, a 54% loss of vegetable production, and a 32% loss of cereal production. Therefore, pesticides play a critical role in reducing diseases and increasing crop yields worldwide. Thus, it is essential to discuss the agricultural development process; the historical perspective, types and specific uses of pesticides; and pesticide behavior, its contamination, and adverse effects on the natural environment. The review study indicates that agricultural development has a long history in many places around the world. The history of pesticide use can be divided into three periods of time. Pesticides are classified by different classification terms such as chemical classes, functional groups, modes of action, and toxicity. Pesticides are used to kill pests and control weeds using chemical ingredients; hence, they can also be toxic to other organisms, including birds, fish, beneficial insects, and non-target plants, as well as air, water, soil, and crops. Moreover, pesticide contamination moves away from the target plants, resulting in environmental pollution. Such chemical residues impact human health through environmental and food contamination. In addition, climate change-related factors also impact on pesticide application and result in increased pesticide usage and pesticide pollution. Therefore, this review will provide the scientific information necessary for pesticide application and management in the future.

Moynihan R., Sanders S., Michaleff Z.A., Scott A.M., Clark J., To E.J., Jones M., Kitchener E., Fox M., Johansson M., Lang E., Duggan A., Scott I., Albarqouni L.

ObjectivesTo determine the extent and nature of changes in utilisation of healthcare services during COVID-19 pandemic.DesignSystematic review.EligibilityEligible studies compared utilisation of services during COVID-19 pandemic to at least one comparable period in prior years. Services included visits, admissions, diagnostics and therapeutics. Studies were excluded if from single centres or studied only patients with COVID-19.Data sourcesPubMed, Embase, Cochrane COVID-19 Study Register and preprints were searched, without language restrictions, until 10 August, using detailed searches with key concepts including COVID-19, health services and impact.Data analysisRisk of bias was assessed by adapting the Risk of Bias in Non-randomised Studies of Interventions tool, and a Cochrane Effective Practice and Organization of Care tool. Results were analysed using descriptive statistics, graphical figures and narrative synthesis.Outcome measuresPrimary outcome was change in service utilisation between prepandemic and pandemic periods. Secondary outcome was the change in proportions of users of healthcare services with milder or more severe illness (eg, triage scores).Results3097 unique references were identified, and 81 studies across 20 countries included, reporting on >11 million services prepandemic and 6.9 million during pandemic. For the primary outcome, there were 143 estimates of changes, with a median 37% reduction in services overall (IQR −51% to −20%), comprising median reductions for visits of 42% (−53% to −32%), admissions 28% (−40% to −17%), diagnostics 31% (−53% to −24%) and for therapeutics 30% (−57% to −19%). Among 35 studies reporting secondary outcomes, there were 60 estimates, with 27 (45%) reporting larger reductions in utilisation among people with a milder spectrum of illness, and 33 (55%) reporting no difference.ConclusionsHealthcare utilisation decreased by about a third during the pandemic, with considerable variation, and with greater reductions among people with less severe illness. While addressing unmet need remains a priority, studies of health impacts of reductions may help health systems reduce unnecessary care in the postpandemic recovery.PROSPERO registration numberCRD42020203729.

Hashim F.A., Houssein E.H., Mabrouk M.S., Al-Atabany W., Mirjalili S.

Several metaheuristic optimization algorithms have been developed to solve the real-world problems recently. This paper proposes a novel metaheuristic algorithm named Henry gas solubility optimization (HGSO), which mimics the behavior governed by Henry’s law to solve challenging optimization problems. Henry’s law is an essential gas law relating the amount of a given gas that is dissolved to a given type and volume of liquid at a fixed temperature. The HGSO algorithm imitates the huddling behavior of gas to balance exploitation and exploration in the search space and avoid local optima. The performance of HGSO is tested on 47 benchmark functions, CEC’17 test suite, and three real-world optimization problems. The results are compared with seven well-known algorithms; the particle swarm optimization (PSO), gravitational search algorithm (GSA), cuckoo search algorithm (CS), grey wolf optimizer (GWO), whale optimization algorithm (WOA), elephant herding algorithm (EHO) and simulated annealing (SA). Additionally, to assess the pairwise statistical performance of the competitive algorithms, a Wilcoxon rank sum test is conducted. The experimental results revealed that HGSO provides competitive and superior results compared to other algorithms when solving challenging optimization problems.

Smolen J.S., Landewé R.B., Bergstra S.A., Kerschbaumer A., Sepriano A., Aletaha D., Caporali R., Edwards C.J., Hyrich K.L., Pope J.E., de Souza S., Stamm T.A., Takeuchi T., Verschueren P., Winthrop K.L., et. al.

ObjectivesTo provide an update of the EULAR rheumatoid arthritis (RA) management recommendations addressing the most recent developments in the field.MethodsAn international task force was formed and solicited three systematic literature research activities on safety and efficacy of disease-modifying antirheumatic drugs (DMARDs) and glucocorticoids (GCs). The new evidence was discussed in light of the last update from 2019. A predefined voting process was applied to each overarching principle and recommendation. Levels of evidence and strengths of recommendation were assigned to and participants finally voted on the level of agreement with each item.ResultsThe task force agreed on 5 overarching principles and 11 recommendations concerning use of conventional synthetic (cs) DMARDs (methotrexate (MTX), leflunomide, sulfasalazine); GCs; biological (b) DMARDs (tumour necrosis factor inhibitors (adalimumab, certolizumab pegol, etanercept, golimumab, infliximab including biosimilars), abatacept, rituximab, tocilizumab, sarilumab and targeted synthetic (ts) DMARDs, namely the Janus kinase inhibitors tofacitinib, baricitinib, filgotinib, upadacitinib. Guidance on monotherapy, combination therapy, treatment strategies (treat-to-target) and tapering in sustained clinical remission is provided. Safety aspects, including risk of major cardiovascular events (MACEs) and malignancies, costs and sequencing of b/tsDMARDs were all considered. Initially, MTX plus GCs is recommended and on insufficient response to this therapy within 3–6 months, treatment should be based on stratification according to risk factors; With poor prognostic factors (presence of autoantibodies, high disease activity, early erosions or failure of two csDMARDs), any bDMARD should be added to the csDMARD; after careful consideration of risks of MACEs, malignancies and/or thromboembolic events tsDMARDs may also be considered in this phase. If the first bDMARD (or tsDMARD) fails, any other bDMARD (from another or the same class) or tsDMARD (considering risks) is recommended. With sustained remission, DMARDs may be tapered but should not be stopped. Levels of evidence and levels of agreement were high for most recommendations.ConclusionsThese updated EULAR recommendations provide consensus on RA management including safety, effectiveness and cost.

Young L., Wright M., Tockner K., Thieme M., Tharme R.E., Robinson J., Ormerod S.J., Olden J.D., Muruven D., McClain M.E., Leonard P., Lynch A.J., Leclère D., Jones T., Hughes K., et. al.

AbstractDespite their limited spatial extent, freshwater ecosystems host remarkable biodiversity, including one-third of all vertebrate species. This biodiversity is declining dramatically: Globally, wetlands are vanishing three times faster than forests, and freshwater vertebrate populations have fallen more than twice as steeply as terrestrial or marine populations. Threats to freshwater biodiversity are well documented but coordinated action to reverse the decline is lacking. We present an Emergency Recovery Plan to bend the curve of freshwater biodiversity loss. Priority actions include accelerating implementation of environmental flows; improving water quality; protecting and restoring critical habitats; managing the exploitation of freshwater ecosystem resources, especially species and riverine aggregates; preventing and controlling nonnative species invasions; and safeguarding and restoring river connectivity. We recommend adjustments to targets and indicators for the Convention on Biological Diversity and the Sustainable Development Goals and roles for national and international state and nonstate actors.

Gursoy D., Chi O.H., Lu L., Nunkoo R.

This study develops and empirically tests a theoretical model of artificially intelligent (AI) device use acceptance (AIDUA) that aims to explain customers’ willingness to accept AI device use in service encounters. The proposed model incorporates three acceptance generation stages (primary appraisal, secondary appraisal, and outcome stage) and six antecedents (social influence, hedonic motivation, anthropomorphism, performance expectancy, effort expectancy, and emotion). Utilizing data collected from potential customers, the proposed AIDUA model is tested. Findings suggest that customers go through a three-step acceptance generation process in determining whether to accept the use of AI devices during their service interactions. Findings indicate that social influence and hedonic motivation are positively related to performance expectancy while anthropomorphism is positively related to effort expectancy. Both performance and effort expectancy are significant antecedents of customer emotions, which determines customers’ acceptance of AI device use in service encounters. This study provides a conceptual AI device acceptance framework that can be used by other researchers to better investigate AI related topics in the service context.

Hirscher M., Yartys V.A., Baricco M., Bellosta von Colbe J., Blanchard D., Bowman R.C., Broom D.P., Buckley C.E., Chang F., Chen P., Cho Y.W., Crivello J., Cuevas F., David W.I., de Jongh P.E., et. al.

Magnesium hydride owns the largest share of publications on solid materials for hydrogen storage. The Magnesium group of international experts contributing to IEA Task 32 Hydrogen Based Energy Storage recently published two review papers presenting the activities of the group focused on magnesium hydride based materials and on Mg based compounds for hydrogen and energy storage. This review article not only overviews the latest activities on both fundamental aspects of Mg-based hydrides and their applications, but also presents a historic overview on the topic and outlines projected future developments. Particular attention is paid to the theoretical and experimental studies of Mg-H system at extreme pressures, kinetics and thermodynamics of the systems based on MgH2,nanostructuring, new Mg-based compounds and novel composites, and catalysis in the Mg based H storage systems. Finally, thermal energy storage and upscaled H storage systems accommodating MgH2 are presented.

Suryanto B.H., Wang Y., Hocking R.K., Adamson W., Zhao C.

Efficient generation of hydrogen from water-splitting is an underpinning chemistry to realize the hydrogen economy. Low cost, transition metals such as nickel and iron-based oxides/hydroxides have been regarded as promising catalysts for the oxygen evolution reaction in alkaline media with overpotentials as low as ~200 mV to achieve 10 mA cm−2, however, they are generally unsuitable for the hydrogen evolution reaction. Herein, we show a Janus nanoparticle catalyst with a nickel–iron oxide interface and multi-site functionality for a highly efficient hydrogen evolution reaction with a comparable performance to the benchmark platinum on carbon catalyst. Density functional theory calculations reveal that the hydrogen evolution reaction catalytic activity of the nanoparticle is induced by the strong electronic coupling effect between the iron oxide and the nickel at the interface. Remarkably, the catalyst also exhibits extraordinary oxygen evolution reaction activity, enabling an active and stable bi-functional catalyst for whole cell water-splitting with, to the best of our knowledge, the highest energy efficiency (83.7%) reported to date. Ni–Fe based compound are known as active electrocatalysts for oxygen evolution reaction, but not a good choice for the other half-reaction of water-splitting. Here the authors report a unique interface between Ni and γ-Fe2O3 that efficiently catalyzes the cathodic hydrogen evolution reaction.