Explore the vast range of titles available.

Practical electrochemical carbon dioxide (CO₂) conversion requires a catalyst capable of mediating the efficient formation of a single product with high selectivity at high current densities. Solid-state electrocatalysts achieve the CO₂ reduction reaction (CO₂RR) at current densities ≥ 150 milliamperes per square centimeter (mA/cm²), but maintaining high selectivities at high current densities and efficiencies remains a challenge. Molecular CO₂RR catalysts can be designed to achieve high selectivities and low overpotentials but only at current densities irrelevant to commercial operation. We show here that cobalt phthalocyanine, a widely available molecular catalyst, can mediate CO₂ to CO formation in a zero-gap membrane flow reactor with selectivities > 95% at 150 mA/cm². The revelation that molecular catalysts can work efficiently under these operating conditions illuminates a distinct approach for optimizing CO₂RR catalysts and electrolyzers.

Based on the change of information transmission form and interactive mode, this paper will introduce the concept of scene-based precise communication in the network live broadcast Based on the analysis of the current situation of the network live broadcast platform, this paper focuses on the construction of the network live broadcast scene from the aspects of technology, content and users The ultimate goal of scene analysis is to understand the logic and habits of users' behavior, and to better serve the audience.

Quantum mechanics governs the microscopic world, where low mass and momentum reveal a natural wave–particle duality. Magnifying quantum behaviour to macroscopic scales is a major strength of the technique of cooling and trapping atomic gases, in which low momentum is engineered through extremely low temperatures. Advances in this field have achieved such precise control over atomic systems that gravity, often negligible when considering individual atoms, has emerged as a substantial obstacle. In particular, although weaker trapping fields would allow access to lower temperatures 1 , 2 , gravity empties atom traps that are too weak. Additionally, inertial sensors based on cold atoms could reach better sensitivities if the free-fall time of the atoms after release from the trap could be made longer 3 . Planetary orbit, specifically the condition of perpetual free-fall, offers to lift cold-atom studies beyond such terrestrial limitations. Here we report production of rubidium Bose–Einstein condensates (BECs) in an Earth-orbiting research laboratory, the Cold Atom Lab. We observe subnanokelvin BECs in weak trapping potentials with free-expansion times extending beyond one second, providing an initial demonstration of the advantages offered by a microgravity environment for cold-atom experiments and verifying the successful operation of this facility. With routine BEC production, continuing operations will support long-term investigations of trap topologies unique to microgravity 4 , 5 , atom-laser sources 6 , few-body physics 7 , 8 and pathfinding techniques for atom-wave interferometry 9 – 12 . A Bose–Einstein condensate prepared in low Earth orbit shows a free-expansion time greater than one second, demonstrating the advantages of a microgravity environment for such studies.

We study the problem of training deep fully connected neural networks with Rectified Linear Unit (ReLU) activation function and cross entropy loss function for binary classification using gradient descent. We show that with proper random weight initialization, gradient descent can find the global minima of the training loss for an over-parameterized deep ReLU network, under certain assumption on the training data. The key idea of our proof is that Gaussian random initialization followed by gradient descent produces a sequence of iterates that stay inside a small perturbation region centered at the initial weights, in which the training loss function of the deep ReLU networks enjoys nice local curvature properties that ensure the global convergence of gradient descent. At the core of our proof technique is (1) a milder assumption on the training data; (2) a sharp analysis of the trajectory length for gradient descent; and (3) a finer characterization of the size of the perturbation region. Compared with the concurrent work (Allen-Zhu et al. in A convergence theory for deep learning via over-parameterization, 2018a; Du et al. in Gradient descent finds global minima of deep neural networks, 2018a) along this line, our result relies on milder over-parameterization condition on the neural network width, and enjoys faster global convergence rate of gradient descent for training deep neural networks.

TWO OXEN AHEAD This revealing study of farming practices in societies around the Mediterranean draws out the valuable contribution that knowledge of recent practices can make to our understanding of husbandry in prehistoric and Greco-Roman times. It reflects increased academic interest in the formative influence of farming regimes on the societies they were designed to feed. The author's intensive research took him to farming communities around the Mediterranean, where he recorded observational and interview data on differing farming strategies and practices, many of which can be traced back to classical antiquity or earlier. The book documents these variables, through the annual chaîne opératoire (from ploughing and sowing to harvesting and threshing), interannual schemes of crop rotation and husbandry, and the generational cycle of household development. It traces the interdependence of these successive stages and explores how cultural tradition, ecological conditions, and access to resources shape variability in husbandry practice. Each chapter identifies ways in which heuristic use of data on recent farming can shed light on ancient practices and societies.

Military named entity recognition is the basis of the military intelligence analysis and operational information service. In order to solve the problems of inaccurate word segmentation, diverse forms and the lack of corpus in military texts, the author proposes a method of military named entity recognition based on Pre-training language model. On this basis, and taking advantage of Bi-directional Long Short-Term Memory (BiLSTM) neural network in dealing with the wide range of contextual information, the BERT-BiLSTM-CRF named entity recognition model was constructed. The experimental results on the tagged military text corpus show that the extraction effect of this method is better than that of the traditional methods.

During epidemics, the medical working environment is highly stressful especially for the nurses. The purpose of this study was to assess occupational stress, job satisfaction, and intent to leave among nurses dealing with suspected COVID-19 patients. A comparative cross-sectional study was conducted among 210 nurses from Zagazig Fever Hospital (ZFH) which is one of COVID-19 Triage Hospitals (Group I) versus 210 nurses from Zagazig General Hospital (ZGH) (Group II) which is neither triage nor isolation hospital; dealing only with suspected COVID-19 patients in emergency at Sharkia Governorate, Egypt, from 10th to 24th of April 2020. Assessment was done through online questionnaire formed of the Expanded Nursing Stress Scale, the McCloskey/Mueller Satisfaction Scale, and questionnaire assessing specific COVID-19-associated stressors and nurses’ intent to leave. Three quarters of nurses (75.2%) in ZFH had high stress level versus 60.5% in ZGH. Workload (98.6%), dealing with death and dying (96.7%), personal demands and fears (95.7%), employing strict biosecurity measures (95.2%), and stigma (90.5%) represented the highest priority stressors in ZFH, while exposure to infection risk (97.6%) was the stressor of highest priority among ZGH according to Pareto analysis. More than half of nurses (51.0%) in ZFH reported low satisfaction level versus 41.9% in ZGH. Only 4.8% of nurses in ZFH definitely had no intent to leave their present job. Type of hospital and its related workload were the most significant predictor of all the studied outcomes.

In view of the complexity and flexibility of the current power grid, in order to make relay protection online verification more accurate and effective, a protection importance degree sequence assessment method based on the multi-source information fusion technology and bowtie model is proposed. Firstly, the cause of incorrect protection actions is analyzed by establishing a fault tree through multi-source information fusion technology, and the event tree analysis method is used to deduce the final development result of the event. Combined with the fault tree and the event tree, a bow-tie model is formed to visually and completely describe the whole process of the event. Then the economic loss of the event is analyzed as the basis for assessing the protection importance degree, and the online inspection of the protection setting value is completed from high to low according to the importance degree sequence. Finally, the IEEE 3-machine 9-bus system is used to verify the correctness and rationality of the method through simulation.

The area of offline handwritten text recognition(OHTR) has been widely researched in the last decades, but it stills an important research problem. The OHTR system has an objective to transform a document image into text data. Compared with online handwriting recognition, the dynamic information about the writing trajectories in OHTR is not available. Many advancements have been proposed in the literature, most notably the application of deep learning methods to OHTR. In this paper, we introduced how this problem has been handled in the past few decades, analyze the latest advancements and the potential directions for future research in this field.

We study the role of establishment-specific wage premiums in generating recent increases in West German wage inequality. Models with additive fixed effects for workers and establishments are fit into four subintervals spanning the period from 1985 to 2009. We show that these models provide a good approximation to the wage structure and can explain nearly all of the dramatic rise in West German wage inequality. Our estimates suggest that the increasing dispersion of West German wages has arisen from a combination of rising heterogeneity between workers, rising dispersion in the wage premiums at different establishments, and increasing assortativeness in the assignment of workers to plants. In contrast, the idiosyncratic job-match component of wage variation is small and stable over time. Decomposing changes in mean wages between different education groups, occupations, and industries, we find that increasing plant-level heterogeneity and rising assortativeness in the assignment of workers to establishments explain a large share of the rise in inequality along all three dimensions.