Abstract: A major challenge today and into the future is to maintain or enhance beneficial contributions of nature to a good quality of life for all people. This is among the key motivations of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES), a joint global effort by governments, academia, and civil society to assess and promote knowledge of Earth's biodiversity and ecosystems and their contribution to human societies in order to inform policy formulation. One of the more recent key elements of the IPBES conceptual framework ( 1 ) is the notion of nature's contributions to people (NCP), which builds on the ecosystem service concept popularized by the Millennium Ecosystem Assessment (MA) ( 2 ). But as we detail below, NCP as defined and put into practice in IPBES differs from earlier work in several important ways. First, the NCP approach recognizes the central and pervasive role that culture plays in defining all links between people and nature. Second, use of NCP elevates, emphasizes, and operationalizes the role of indigenous and local knowledge in understanding nature's contribution to people.

Abstract: Convolutional Neural Networks (CNNs) have proven very effective in image classification and show promise for audio. We use various CNN architectures to classify the soundtracks of a dataset of 70M training videos (5.24 million hours) with 30,871 video-level labels. We examine fully connected Deep Neural Networks (DNNs), AlexNet [1], VGG [2], Inception [3], and ResNet [4]. We investigate varying the size of both training set and label vocabulary, finding that analogs of the CNNs used in image classification do well on our audio classification task, and larger training and label sets help up to a point. A model using embeddings from these classifiers does much better than raw features on the Audio Set [5] Acoustic Event Detection (AED) classification task.

Abstract: Fungal diseases kill more than 1.5 million and affect over a billion people. However, they are still a neglected topic by public health authorities even though most deaths from fungal diseases are avoidable. Serious fungal infections occur as a consequence of other health problems including asthma, AIDS, cancer, organ transplantation and corticosteroid therapies. Early accurate diagnosis allows prompt antifungal therapy; however this is often delayed or unavailable leading to death, serious chronic illness or blindness. Recent global estimates have found 3,000,000 cases of chronic pulmonary aspergillosis, ~223,100 cases of cryptococcal meningitis complicating HIV/AIDS, ~700,000 cases of invasive candidiasis, ~500,000 cases of Pneumocystis jirovecii pneumonia, ~250,000 cases of invasive aspergillosis, ~100,000 cases of disseminated histoplasmosis, over 10,000,000 cases of fungal asthma and ~1,000,000 cases of fungal keratitis occur annually. Since 2013, the Leading International Fungal Education (LIFE) portal has facilitated the estimation of the burden of serious fungal infections country by country for over 5.7 billion people (>80% of the world’s population). These studies have shown differences in the global burden between countries, within regions of the same country and between at risk populations. Here we interrogate the accuracy of these fungal infection burden estimates in the 43 published papers within the LIFE initiative.

Abstract: Convolutional neural network-based approaches for semantic segmentation rely on supervision with pixel-level ground truth, but may not generalize well to unseen image domains. As the labeling process is tedious and labor intensive, developing algorithms that can adapt source ground truth labels to the target domain is of great interest. In this paper, we propose an adversarial learning method for domain adaptation in the context of semantic segmentation. Considering semantic segmentations as structured outputs that contain spatial similarities between the source and target domains, we adopt adversarial learning in the output space. To further enhance the adapted model, we construct a multi-level adversarial network to effectively perform output space domain adaptation at different feature levels. Extensive experiments and ablation study are conducted under various domain adaptation settings, including synthetic-to-real and cross-city scenarios. We show that the proposed method performs favorably against the state-of-the-art methods in terms of accuracy and visual quality.

Abstract: Deep Learning has revolutionized vision via convolutional neural networks (CNNs) and natural language processing via recurrent neural networks (RNNs). However, success stories of Deep Learning with standard feed-forward neural networks (FNNs) are rare. FNNs that perform well are typically shallow and, therefore cannot exploit many levels of abstract representations. We introduce self-normalizing neural networks (SNNs) to enable high-level abstract representations. While batch normalization requires explicit normalization, neuron activations of SNNs automatically converge towards zero mean and unit variance. The activation function of SNNs are "scaled exponential linear units" (SELUs), which induce self-normalizing properties. Using the Banach fixed-point theorem, we prove that activations close to zero mean and unit variance that are propagated through many network layers will converge towards zero mean and unit variance -- even under the presence of noise and perturbations. This convergence property of SNNs allows to (1) train deep networks with many layers, (2) employ strong regularization, and (3) to make learning highly robust. Furthermore, for activations not close to unit variance, we prove an upper and lower bound on the variance, thus, vanishing and exploding gradients are impossible. We compared SNNs on (a) 121 tasks from the UCI machine learning repository, on (b) drug discovery benchmarks, and on (c) astronomy tasks with standard FNNs and other machine learning methods such as random forests and support vector machines. SNNs significantly outperformed all competing FNN methods at 121 UCI tasks, outperformed all competing methods at the Tox21 dataset, and set a new record at an astronomy data set. The winning SNN architectures are often very deep. Implementations are available at: this http URL.

Abstract: INTRODUCTION Dendritic cells (DCs) and monocytes consist of multiple specialized subtypes that play a central role in pathogen sensing, phagocytosis, and antigen presentation. However, their identities and interrelationships are not fully understood, as these populations have historically been defined by a combination of morphology, physical properties, localization, functions, developmental origins, and expression of a restricted set of surface markers. RATIONALE To overcome this inherently biased strategy for cell identification, we performed single-cell RNA sequencing of ~2400 cells isolated from healthy blood donors and enriched for HLA-DR + lineage − cells. This single-cell profiling strategy and unbiased genomic classification, together with follow-up profiling and functional and phenotypic characterization of prospectively isolated subsets, led us to identify and validate six DC subtypes and four monocyte subtypes, and thus revise the taxonomy of these cells. RESULTS Our study reveals: 1) A new DC subset, representing 2 to 3% of the DC populations across all 10 donors tested, characterized by the expression of AXL , SIGLEC1 , and SIGLEC6 antigens, named AS DCs. The AS DC population further divides into two populations captured in the traditionally defined plasmacytoid DC (pDC) and CD1C + conventional DC (cDC) gates. This split is further reflected through AS DC gene expression signatures spanning a spectrum between cDC-like and pDC-like gene sets. Although AS DCs share properties with pDCs, they more potently activate T cells. This discovery led us to reclassify pDCs as the originally described “natural interferon-producing cells (IPCs)” with weaker T cell proliferation induction ability. 2) A new subdivision within the CD1C + DC subset: one defined by a major histocompatibility complex class II–like gene set and one by a CD14 + monocyte–like prominent gene set. These CD1C + DC subsets, which can be enriched by combining CD1C with CD32B, CD36, and CD163 antigens, can both potently induce T cell proliferation. 3) The existence of a circulating and dividing cDC progenitor giving rise to CD1C + and CLEC9A + DCs through in vitro differentiation assays. This blood precursor is defined by the expression of CD100 + CD34 int and observed at a frequency of ~0.02% of the LIN – HLA-DR + fraction. 4) Two additional monocyte populations: one expressing classical monocyte genes and cytotoxic genes, and the other with unknown functions. 5) Evidence for a relationship between blastic plasmacytoid DC neoplasia (BPDCN) cells and healthy DCs. CONCLUSION Our revised taxonomy will enable more accurate functional and developmental analyses as well as immune monitoring in health and disease. The discovery of AS DCs within the traditionally defined pDC population explains many of the cDC properties previously assigned to pDCs, highlighting the need to revisit the definition of pDCs. Furthermore, the discovery of blood cDC progenitors represents a new therapeutic target readily accessible in the bloodstream for manipulation, as well as a new source for better in vitro DC generation. Although the current results focus on DCs and monocytes, a similar strategy can be applied to build a comprehensive human immune cell atlas.

Abstract: During the 2008-2009 financial crisis, firms with high social capital, measured as corporate social responsibility (CSR) intensity, had stock returns that were four to seven percentage points higher than firms with low social capital. High-CSR firms also experienced higher profitability, growth, and sales per employee relative to low-CSR firms, and they raised more debt. This evidence suggests that the trust between the firm and both its stakeholders and investors, built through investments in social capital, pays off when the overall level of trust in corporations and markets suffers a negative shock.

Abstract: Importance Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes coronavirus disease 2019 (COVID-19) has reached a pandemic level. Coronaviruses are known to affect the cardiovascular system. We review the basics of coronaviruses, with a focus on COVID-19, along with their effects on the cardiovascular system. Observations Coronavirus disease 2019 can cause a viral pneumonia with additional extrapulmonary manifestations and complications. A large proportion of patients have underlying cardiovascular disease and/or cardiac risk factors. Factors associated with mortality include male sex, advanced age, and presence of comorbidities including hypertension, diabetes mellitus, cardiovascular diseases, and cerebrovascular diseases. Acute cardiac injury determined by elevated high-sensitivity troponin levels is commonly observed in severe cases and is strongly associated with mortality. Acute respiratory distress syndrome is also strongly associated with mortality. Conclusions and Relevance Coronavirus disease 2019 is associated with a high inflammatory burden that can induce vascular inflammation, myocarditis, and cardiac arrhythmias. Extensive efforts are underway to find specific vaccines and antivirals against SARS-CoV-2. Meanwhile, cardiovascular risk factors and conditions should be judiciously controlled per evidence-based guidelines.

Abstract: Resena de la obra de Don A. Dillman, Jolene D. Smyth y Leah Melani Christian: Internet, Phone, Mail and Mixed-Mode Surveys. The Tailored Design Method. New Jersey: John Wiley and Sons

Abstract: Summary Background Palbociclib (PD-0332991) is an oral, small-molecule inhibitor of cyclin-dependent kinases (CDKs) 4 and 6 with preclinical evidence of growth-inhibitory activity in oestrogen receptor-positive breast cancer cells and synergy with anti-oestrogens. We aimed to assess the safety and efficacy of palbociclib in combination with letrozole as first-line treatment of patients with advanced, oestrogen receptor-positive, HER2-negative breast cancer. Methods In this open-label, randomised phase 2 study, postmenopausal women with advanced oestrogen receptor-positive and HER2-negative breast cancer who had not received any systemic treatment for their advanced disease were eligible to participate. Patients were enrolled in two separate cohorts that accrued sequentially: in cohort 1, patients were enrolled on the basis of their oestrogen receptor-positive and HER2-negative biomarker status alone, whereas in cohort 2 they were also required to have cancers with amplification of cyclin D1 ( CCND1 ), loss of p16 (INK4A or CDKN2A), or both. In both cohorts, patients were randomly assigned 1:1 via an interactive web-based randomisation system, stratified by disease site and disease-free interval, to receive continuous oral letrozole 2·5 mg daily or continuous oral letrozole 2·5 mg daily plus oral palbociclib 125 mg, given once daily for 3 weeks followed by 1 week off over 28-day cycles. The primary endpoint was investigator-assessed progression-free survival in the intention-to-treat population. Accrual to cohort 2 was stopped after an unplanned interim analysis of cohort 1 and the statistical analysis plan for the primary endpoint was amended to a combined analysis of cohorts 1 and 2 (instead of cohort 2 alone). The study is ongoing but closed to accrual; these are the results of the final analysis of progression-free survival. The study is registered with the ClinicalTrials.gov, number NCT00721409. Findings Between Dec 22, 2009, and May 12, 2012, we randomly assigned 165 patients, 84 to palbociclib plus letrozole and 81 to letrozole alone. At the time of the final analysis for progression-free survival (median follow-up 29·6 months [95% CI 27·9–36·0] for the palbociclib plus letrozole group and 27·9 months [25·5–31·1] for the letrozole group), 41 progression-free survival events had occurred in the palbociclib plus letrozole group and 59 in the letrozole group. Median progression-free survival was 10·2 months (95% CI 5·7–12·6) for the letrozole group and 20·2 months (13·8–27·5) for the palbociclib plus letrozole group (HR 0·488, 95% CI 0·319–0·748; one-sided p=0·0004). In cohort 1 (n=66), median progression-free survival was 5·7 months (2·6–10·5) for the letrozole group and 26·1 months (11·2–not estimable) for the palbociclib plus letrozole group (HR 0·299, 0·156–0·572; one-sided p Interpretation The addition of palbociclib to letrozole in this phase 2 study significantly improved progression-free survival in women with advanced oestrogen receptor-positive and HER2-negative breast cancer. A phase 3 trial is currently underway. Funding Pfizer.

Abstract: BACKGROUND New treatments have improved outcomes for patients with relapsed chronic lymphocytic leukemia (CLL), but complete remissions remain uncommon. Venetoclax has a distinct mechanism of action; it targets BCL2, a protein central to the survival of CLL cells. METHODS We conducted a phase 1 dose-escalation study of daily oral venetoclax in patients with relapsed or refractory CLL or small lymphocytic lymphoma (SLL) to assess safety, pharmacokinetic profile, and efficacy. In the dose-escalation phase, 56 patients received active treatment in one of eight dose groups that ranged from 150 to 1200 mg per day. In an expansion cohort, 60 additional patients were treated with a weekly stepwise ramp-up in doses as high as 400 mg per day. RESULTS The majority of the study patients had received multiple previous treatments, and 89% had poor prognostic clinical or genetic features. Venetoclax was active at all dose levels. Clinical tumor lysis syndrome occurred in 3 of 56 patients in the dose-escalation cohort, with one death. After adjustments to the dose-escalation schedule, clinical tumor lysis syndrome did not occur in any of the 60 patients in the expansion cohort. Other toxic effects in cluded mild diarrhea (in 52% of the patients), upper respiratory tract infection (in 48%), nausea (in 47%), and grade 3 or 4 neutropenia (in 41%). A maximum tolerated dose was not identified. Among the 116 patients who received venetoclax, 92 (79%) had a response. Response rates ranged from 71 to 79% among patients in subgroups with an adverse prognosis, including those with resistance to fludarabine, those with chromosome 17p deletions (deletion 17p CLL), and those with unmutated IGHV. Complete remissions occurred in 20% of the patients, including 5% who had no minimal residual disease on flow cytometry. The 15-month progression-free survival estimate for the 400-mg dose groups was 69%. CONCLUSIONS Selective targeting of BCL2 with venetoclax had a manageable safety profile and induced substantial responses in patients with relapsed CLL or SLL, including those with poor prognostic features. (Funded by AbbVie and Genentech; ClinicalTrials.gov number, NCT01328626.)

Abstract: CellProfiler has enabled the scientific research community to create flexible, modular image analysis pipelines since its release in 2005. Here, we describe CellProfiler 3.0, a new version of the software supporting both whole-volume and plane-wise analysis of three-dimensional (3D) image stacks, increasingly common in biomedical research. CellProfiler's infrastructure is greatly improved, and we provide a protocol for cloud-based, large-scale image processing. New plugins enable running pretrained deep learning models on images. Designed by and for biologists, CellProfiler equips researchers with powerful computational tools via a well-documented user interface, empowering biologists in all fields to create quantitative, reproducible image analysis workflows.

Abstract: The IUPHAR/BPS Guide to PHARMACOLOGY (GtoPdb, www.guidetopharmacology.org) and its precursor IUPHAR-DB, have captured expert-curated interactions between targets and ligands from selected papers in pharmacology and drug discovery since 2003. This resource continues to be developed in conjunction with the International Union of Basic and Clinical Pharmacology (IUPHAR) and the British Pharmacological Society (BPS). As previously described, our unique model of content selection and quality control is based on 96 target-class subcommittees comprising 512 scientists collaborating with in-house curators. This update describes content expansion, new features and interoperability improvements introduced in the 10 releases since August 2015. Our relationship matrix now describes ∼9000 ligands, ∼15 000 binding constants, ∼6000 papers and ∼1700 human proteins. As an important addition, we also introduce our newly funded project for the Guide to IMMUNOPHARMACOLOGY (GtoImmuPdb, www.guidetoimmunopharmacology.org). This has been 'forked' from the well-established GtoPdb data model and expanded into new types of data related to the immune system and inflammatory processes. This includes new ligands, targets, pathways, cell types and diseases for which we are recruiting new IUPHAR expert committees. Designed as an immunopharmacological gateway, it also has an emphasis on potential therapeutic interventions.

Abstract: Machine learning is enabling a myriad innovations, including new algorithms for cancer diagnosis and self-driving cars. The broad use of machine learning makes it important to understand the extent to which machine-learning algorithms are subject to attack, particularly when used in applications where physical security or safety is at risk. In this paper, we focus on facial biometric systems, which are widely used in surveillance and access control. We define and investigate a novel class of attacks: attacks that are physically realizable and inconspicuous, and allow an attacker to evade recognition or impersonate another individual. We develop a systematic method to automatically generate such attacks, which are realized through printing a pair of eyeglass frames. When worn by the attacker whose image is supplied to a state-of-the-art face-recognition algorithm, the eyeglasses allow her to evade being recognized or to impersonate another individual. Our investigation focuses on white-box face-recognition systems, but we also demonstrate how similar techniques can be used in black-box scenarios, as well as to avoid face detection.

Abstract: mTORC1 is an important regulator of muscle mass but how it is modulated by oxygen and nutrients is not completely understood. We show that loss of the prolyl hydroxylase domain isoform 1 oxygen sensor in mice (PHD1KO) reduces muscle mass. PHD1KO muscles show impaired mTORC1 activation in response to leucine whereas mTORC1 activation by growth factors or eccentric contractions was preserved. The ability of PHD1 to promote mTORC1 activity is independent of its hydroxylation activity but is caused by decreased protein content of the leucyl tRNA synthetase (LRS) leucine sensor. Mechanistically, PHD1 interacts with and stabilizes LRS. This interaction is promoted during oxygen and amino acid depletion and protects LRS from degradation. Finally, elderly subjects have lower PHD1 levels and LRS activity in muscle from aged versus young human subjects. In conclusion, PHD1 ensures an optimal mTORC1 response to leucine after episodes of metabolic scarcity.

Abstract: Semi-supervised learning has proven to be a powerful paradigm for leveraging unlabeled data to mitigate the reliance on large labeled datasets. In this work, we unify the current dominant approaches for semi-supervised learning to produce a new algorithm, MixMatch, that works by guessing low-entropy labels for data-augmented unlabeled examples and mixing labeled and unlabeled data using MixUp. We show that MixMatch obtains state-of-the-art results by a large margin across many datasets and labeled data amounts. For example, on CIFAR-10 with 250 labels, we reduce error rate by a factor of 4 (from 38% to 11%) and by a factor of 2 on STL-10. We also demonstrate how MixMatch can help achieve a dramatically better accuracy-privacy trade-off for differential privacy. Finally, we perform an ablation study to tease apart which components of MixMatch are most important for its success.

Abstract: We report a method to convert discrete representations of molecules to and from a multidimensional continuous representation. This model allows us to generate new molecules for efficient exploration and optimization through open-ended spaces of chemical compounds. A deep neural network was trained on hundreds of thousands of existing chemical structures to construct three coupled functions: an encoder, a decoder and a predictor. The encoder converts the discrete representation of a molecule into a real-valued continuous vector, and the decoder converts these continuous vectors back to discrete molecular representations. The predictor estimates chemical properties from the latent continuous vector representation of the molecule. Continuous representations allow us to automatically generate novel chemical structures by performing simple operations in the latent space, such as decoding random vectors, perturbing known chemical structures, or interpolating between molecules. Continuous representations also allow the use of powerful gradient-based optimization to efficiently guide the search for optimized functional compounds. We demonstrate our method in the domain of drug-like molecules and also in the set of molecules with fewer that nine heavy atoms.

Abstract: Deep convolutional networks have become a popular tool for image generation and restoration. Generally, their excellent performance is imputed to their ability to learn realistic image priors from a large number of example images. In this paper, we show that, on the contrary, the structure of a generator network is sufficient to capture a great deal of low-level image statistics prior to any learning. In order to do so, we show that a randomly-initialized neural network can be used as a handcrafted prior with excellent results in standard inverse problems such as denoising, superresolution, and inpainting. Furthermore, the same prior can be used to invert deep neural representations to diagnose them, and to restore images based on flash-no flash input pairs. Apart from its diverse applications, our approach highlights the inductive bias captured by standard generator network architectures. It also bridges the gap between two very popular families of image restoration methods: learning-based methods using deep convolutional networks and learning-free methods based on handcrafted image priors such as self-similarity.

Abstract: Reactive oxygen species (ROS) play a key role in the acclimation process of plants to abiotic stress. They primarily function as signal transduction molecules that regulate different pathways during plant acclimation to stress, but are also toxic byproducts of stress metabolism. Because each subcellular compartment in plants contains its own set of ROS-producing and ROS-scavenging pathways, the steady-state level of ROS, as well as the redox state of each compartment, is different at any given time giving rise to a distinct signature of ROS levels at the different compartments of the cell. Here we review recent studies on the role of ROS in abiotic stress in plants, and propose that different abiotic stresses, such as drought, heat, salinity and high light, result in different ROS signatures that determine the specificity of the acclimation response and help tailor it to the exact stress the plant encounters. We further address the role of ROS in the acclimation of plants to stress combination as well as the role of ROS in mediating rapid systemic signaling during abiotic stress. We conclude that as long as cells maintain high enough energy reserves to detoxify ROS, ROS is beneficial to plants during abiotic stress enabling them to adjust their metabolism and mount a proper acclimation response.

Abstract: We describe Human Mesh Recovery (HMR), an end-to-end framework for reconstructing a full 3D mesh of a human body from a single RGB image. In contrast to most current methods that compute 2D or 3D joint locations, we produce a richer and more useful mesh representation that is parameterized by shape and 3D joint angles. The main objective is to minimize the reprojection loss of keypoints, which allows our model to be trained using in-the-wild images that only have ground truth 2D annotations. However, the reprojection loss alone is highly underconstrained. In this work we address this problem by introducing an adversary trained to tell whether human body shape and pose parameters are real or not using a large database of 3D human meshes. We show that HMR can be trained with and without using any paired 2D-to-3D supervision. We do not rely on intermediate 2D keypoint detections and infer 3D pose and shape parameters directly from image pixels. Our model runs in real-time given a bounding box containing the person. We demonstrate our approach on various images in-the-wild and out-perform previous optimization-based methods that output 3D meshes and show competitive results on tasks such as 3D joint location estimation and part segmentation.

Abstract: Policies aiming to preserve vegetated coastal ecosystems (VCE; tidal marshes, mangroves and seagrasses) to mitigate greenhouse gas emissions require national assessments of blue carbon resources. Here, we present organic carbon (C) storage in VCE across Australian climate regions and estimate potential annual CO2 emission benefits of VCE conservation and restoration. Australia contributes 5–11% of the C stored in VCE globally (70–185 Tg C in aboveground biomass, and 1,055–1,540 Tg C in the upper 1 m of soils). Potential CO2 emissions from current VCE losses are estimated at 2.1–3.1 Tg CO2-e yr-1, increasing annual CO2 emissions from land use change in Australia by 12–21%. This assessment, the most comprehensive for any nation to-date, demonstrates the potential of conservation and restoration of VCE to underpin national policy development for reducing greenhouse gas emissions. Policies aiming to preserve vegetated coastal ecosystems (VCE) to mitigate greenhouse gas emissions require national assessments of blue carbon resources. Here the authors assessed organic carbon storage in VCE across Australian and the potential annual CO2 emission benefits of VCE conservation and find that Australia contributes substantially the carbon stored in VCE globally.

Abstract: Background Caesarean section (CS) rates continue to evoke worldwide concern because of their steady increase, lack of consensus on the appropriate CS rate and the associated additional short- and long-term risks and costs. We present the latest CS rates and trends over the last 24 years. Methods We collected nationally-representative data on CS rates between 1990 to 2014 and calculated regional and subregional weighted averages. We conducted a longitudinal analysis calculating differences in CS rates as absolute change and as the average annual rate of increase (AARI). Results According to the latest data from 150 countries, currently 18.6% of all births occur by CS, ranging from 6% to 27.2% in the least and most developed regions, respectively. Latin America and the Caribbean region has the highest CS rates (40.5%), followed by Northern America (32.3%), Oceania (31.1%), Europe (25%), Asia (19.2%) and Africa (7.3%). Based on the data from 121 countries, the trend analysis showed that between 1990 and 2014, the global average CS rate increased 12.4% (from 6.7% to 19.1%) with an average annual rate of increase of 4.4%. The largest absolute increases occurred in Latin America and the Caribbean (19.4%, from 22.8% to 42.2%), followed by Asia (15.1%, from 4.4% to 19.5%), Oceania (14.1%, from 18.5% to 32.6%), Europe (13.8%, from 11.2% to 25%), Northern America (10%, from 22.3% to 32.3%) and Africa (4.5%, from 2.9% to 7.4%). Asia and Northern America were the regions with the highest and lowest average annual rate of increase (6.4% and 1.6%, respectively). Conclusion The use of CS worldwide has increased to unprecedented levels although the gap between higher- and lower-resource settings remains. The information presented is essential to inform policy and global and regional strategies aimed at optimizing the use of CS.

Abstract: Understanding human motion behavior is critical for autonomous moving platforms (like self-driving cars and social robots) if they are to navigate human-centric environments. This is challenging because human motion is inherently multimodal: given a history of human motion paths, there are many socially plausible ways that people could move in the future. We tackle this problem by combining tools from sequence prediction and generative adversarial networks: a recurrent sequence-to-sequence model observes motion histories and predicts future behavior, using a novel pooling mechanism to aggregate information across people. We predict socially plausible futures by training adversarially against a recurrent discriminator, and encourage diverse predictions with a novel variety loss. Through experiments on several datasets we demonstrate that our approach outperforms prior work in terms of accuracy, variety, collision avoidance, and computational complexity.

Abstract: . The System for Automated Geoscientific Analyses (SAGA) is an open source geographic information system (GIS), mainly licensed under the GNU General Public License. Since its first release in 2004, SAGA has rapidly developed from a specialized tool for digital terrain analysis to a comprehensive and globally established GIS platform for scientific analysis and modeling. SAGA is coded in C++ in an object oriented design and runs under several operating systems including Windows and Linux. Key functional features of the modular software architecture comprise an application programming interface for the development and implementation of new geoscientific methods, a user friendly graphical user interface with many visualization options, a command line interpreter, and interfaces to interpreted languages like R and Python. The current version 2.1.4 offers more than 600 tools, which are implemented in dynamically loadable libraries or shared objects and represent the broad scopes of SAGA in numerous fields of geoscientific endeavor and beyond. In this paper, we inform about the system's architecture, functionality, and its current state of development and implementation. Furthermore, we highlight the wide spectrum of scientific applications of SAGA in a review of published studies, with special emphasis on the core application areas digital terrain analysis, geomorphology, soil science, climatology and meteorology, as well as remote sensing.

Abstract: Elaborate design of highly active and stable catalysts from Earth-abundant elements has great potential to produce materials that can replace the noble-metal-based catalysts commonly used in a range of useful (electro)chemical processes. Here we report, for the first time, a synthetic method that leads to in situ growth of {210} high-index faceted Ni3S2 nanosheet arrays on nickel foam (NF). We show that the resulting material, denoted Ni3S2/NF, can serve as a highly active, binder-free, bifunctional electrocatalyst for both the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). Ni3S2/NF is found to give ∼100% Faradaic yield toward both HER and OER and to show remarkable catalytic stability (for >200 h). Experimental results and theoretical calculations indicate that Ni3S2/NF’s excellent catalytic activity is mainly due to the synergistic catalytic effects produced in it by its nanosheet arrays and exposed {210} high-index facets.