Abstract: Cationic host defence peptides (CHDP), also known as antimicrobial peptides, are naturally occurring peptides that can combat infections through their direct microbicidal properties and/or by influencing the host's immune responses. The unique ability of CHDP to control infections as well as resolve harmful inflammation has generated interest in harnessing the properties of these peptides to develop new therapies for infectious diseases, chronic inflammatory disorders and wound healing. Various strategies have been used to design synthetic optimized peptides, with negligible toxicity. Here, we focus on the progress made in understanding the scope of functions of CHDP and the emerging potential clinical applications of CHDP-based therapies.

Abstract: Person re-identification (re-ID) models trained on one domain often fail to generalize well to another. In our attempt, we present a "learning via translation" framework. In the baseline, we translate the labeled images from source to target domain in an unsupervised manner. We then train re-ID models with the translated images by supervised methods. Yet, being an essential part of this framework, unsupervised image-image translation suffers from the information loss of source-domain labels during translation. Our motivation is two-fold. First, for each image, the discriminative cues contained in its ID label should be maintained after translation. Second, given the fact that two domains have entirely different persons, a translated image should be dissimilar to any of the target IDs. To this end, we propose to preserve two types of unsupervised similarities, 1) self-similarity of an image before and after translation, and 2) domain-dissimilarity of a translated source image and a target image. Both constraints are implemented in the similarity preserving generative adversarial network (SPGAN) which consists of an Siamese network and a CycleGAN. Through domain adaptation experiment, we show that images generated by SPGAN are more suitable for domain adaptation and yield consistent and competitive re-ID accuracy on two large-scale datasets.

Abstract: We systematically explore regularizing neural networks by penalizing low entropy output distributions. We show that penalizing low entropy output distributions, which has been shown to improve exploration in reinforcement learning, acts as a strong regularizer in supervised learning. Furthermore, we connect a maximum entropy based confidence penalty to label smoothing through the direction of the KL divergence. We exhaustively evaluate the proposed confidence penalty and label smoothing on 6 common benchmarks: image classification (MNIST and Cifar-10), language modeling (Penn Treebank), machine translation (WMT'14 English-to-German), and speech recognition (TIMIT and WSJ). We find that both label smoothing and the confidence penalty improve state-of-the-art models across benchmarks without modifying existing hyperparameters, suggesting the wide applicability of these regularizers.

Abstract: As wireless video is the fastest growing form of data traffic, methods for spectrally efficient on-demand wireless video streaming are essential to both service providers and users. A key property of video on-demand is the asynchronous content reuse , such that a few popular files account for a large part of the traffic but are viewed by users at different times. Caching of content on wireless devices in conjunction with device-to-device (D2D) communications allows to exploit this property, and provide a network throughput that is significantly in excess of both the conventional approach of unicasting from cellular base stations and the traditional D2D networks for “regular” data traffic. This paper presents in a tutorial and concise form some recent results on the throughput scaling laws of wireless networks with caching and asynchronous content reuse, contrasting the D2D approach with other alternative approaches such as conventional unicasting, harmonic broadcasting , and a novel coded multicasting approach based on caching in the user devices and network-coded transmission from the cellular base station only. Somehow surprisingly, the D2D scheme with spatial reuse and simple decentralized random caching achieves the same near-optimal throughput scaling law as coded multicasting. Both schemes achieve an unbounded throughput gain (in terms of scaling law) with respect to conventional unicasting and harmonic broadcasting, in the relevant regime where the number of video files in the library is smaller than the total size of the distributed cache capacity in the network. To better understand the relative merits of these competing approaches, we consider a holistic D2D system design incorporating traditional microwave (2 GHz) and millimeter-wave (mm-wave) D2D links; the direct connections to the base station can be used to provide those rare video requests that cannot be found in local caches. We provide extensive simulation results under a variety of system settings and compare our scheme with the systems that exploit transmission from the base station only. We show that, also in realistic conditions and nonasymptotic regimes, the proposed D2D approach offers very significant throughput gains.

Abstract: Current speaker verification techniques rely on a neural network to extract speaker representations. The successful x-vector architecture is a Time Delay Neural Network (TDNN) that applies statistics pooling to project variable-length utterances into fixed-length speaker characterizing embeddings. In this paper, we propose multiple enhancements to this architecture based on recent trends in the related fields of face verification and computer vision. Firstly, the initial frame layers can be restructured into 1-dimensional Res2Net modules with impactful skip connections. Similarly to SE-ResNet, we introduce Squeeze-and-Excitation blocks in these modules to explicitly model channel interdependencies. The SE block expands the temporal context of the frame layer by rescaling the channels according to global properties of the recording. Secondly, neural networks are known to learn hierarchical features, with each layer operating on a different level of complexity. To leverage this complementary information, we aggregate and propagate features of different hierarchical levels. Finally, we improve the statistics pooling module with channel-dependent frame attention. This enables the network to focus on different subsets of frames during each of the channel’s statistics estimation. The proposed ECAPA-TDNN architecture significantly outperforms state-of-the-art TDNN based systems on the VoxCeleb test sets and the 2019 VoxCeleb Speaker Recognition Challenge.

Abstract: The biological effects of cannabinoids, the major constituents of the ancient medicinal plant Cannabis sativa (marijuana) are mediated by two members of the G-protein coupled receptor family, cannabinoid receptors 1 (CB1R) and 2. The CB1R is the prominent subtype in the central nervous system (CNS) and has drawn great attention as a potential therapeutic avenue in several pathological conditions, including neuropsychological disorders and neurodegenerative diseases. Furthermore, cannabinoids also modulate signal transduction pathways and exert profound effects at peripheral sites. Although cannabinoids have therapeutic potential, their psychoactive effects have largely limited their use in clinical practice. In this review, we briefly summarized our knowledge of cannabinoids and the endocannabinoid system, focusing on the CB1R and the CNS, with emphasis on recent breakthroughs in the field. We aim to define several potential roles of cannabinoid receptors in the modulation of signaling pathways and in association with several pathophysiological conditions. We believe that the therapeutic significance of cannabinoids is masked by the adverse effects and here alternative strategies are discussed to take therapeutic advantage of cannabinoids.

Abstract: An intriguing property of deep neural networks is the existence of adversarial examples, which can transfer among different architectures. These transferable adversarial examples may severely hinder deep neural network-based applications. Previous works mostly study the transferability using small scale datasets. In this work, we are the first to conduct an extensive study of the transferability over large models and a large scale dataset, and we are also the first to study the transferability of targeted adversarial examples with their target labels. We study both non-targeted and targeted adversarial examples, and show that while transferable non-targeted adversarial examples are easy to find, targeted adversarial examples generated using existing approaches almost never transfer with their target labels. Therefore, we propose novel ensemble-based approaches to generating transferable adversarial examples. Using such approaches, we observe a large proportion of targeted adversarial examples that are able to transfer with their target labels for the first time. We also present some geometric studies to help understanding the transferable adversarial examples. Finally, we show that the adversarial examples generated using ensemble-based approaches can successfully attack this http URL, which is a black-box image classification system.

Abstract: We develop an algorithm which exceeds the performance of board certified cardiologists in detecting a wide range of heart arrhythmias from electrocardiograms recorded with a single-lead wearable monitor. We build a dataset with more than 500 times the number of unique patients than previously studied corpora. On this dataset, we train a 34-layer convolutional neural network which maps a sequence of ECG samples to a sequence of rhythm classes. Committees of board-certified cardiologists annotate a gold standard test set on which we compare the performance of our model to that of 6 other individual cardiologists. We exceed the average cardiologist performance in both recall (sensitivity) and precision (positive predictive value).

Abstract: This paper presents the results of the WMT16 shared tasks, which included five machine translation (MT) tasks (standard news, IT-domain, biomedical, multimodal, pronoun), three evaluation tasks (metrics, tuning, run-time estimation of MT quality), and an automatic post-editing task and bilingual document alignment task. This year, 102 MT systems from 24 institutions (plus 36 anonymized online systems) were submitted to the 12 translation directions in the news translation task. The IT-domain task received 31 submissions from 12 institutions in 7 directions and the Biomedical task received 15 submissions systems from 5 institutions. Evaluation was both automatic and manual (relative ranking and 100-point scale assessments). The quality estimation task had three subtasks, with a total of 14 teams, submitting 39 entries. The automatic post-editing task had a total of 6 teams, submitting 11 entries.

Abstract: We develop a general approach to valid inference after model selection. At the core of our framework is a result that characterizes the distribution of a post-selection estimator conditioned on the selection event. We specialize the approach to model selection by the lasso to form valid confidence intervals for the selected coefficients and test whether all relevant variables have been included in the model.

Abstract: Transformer-based models have pushed state of the art in many areas of NLP, but our understanding of what is behind their success is still limited. This paper is the first survey of over 150 studies of the popular BERT model. We review the current state of knowledge about how BERT works, what kind of information it learns and how it is represented, common modifications to its training objectives and architecture, the overparameterization issue and approaches to compression. We then outline directions for future research.

Abstract: WEKA is a widely used, open-source machine learning platform. Due to its intuitive interface, it is particularly popular with novice users. However, such users often find it hard to identify the best approach for their particular dataset among the many available. We describe the new version of Auto-WEKA, a system designed to help such users by automatically searching through the joint space of WEKA's learning algorithms and their respective hyperparameter settings to maximize performance, using a state-of-the-art Bayesian optimization method. Our new package is tightly integrated with WEKA, making it just as accessible to end users as any other learning algorithm.

Abstract: In this paper, the biosynthesis process of phenolic compounds in plants is summarized, which includes the shikimate, pentose phosphate and phenylpropanoid pathways. Plant phenolic compounds can act as antioxidants, structural polymers (lignin), attractants (flavonoids and carotenoids), UV screens (flavonoids), signal compounds (salicylic acid and flavonoids) and defense response chemicals (tannins and phytoalexins). From a human physiological standpoint, phenolic compounds are vital in defense responses, such as anti-aging, anti-inflammatory, antioxidant and anti-proliferative activities. Therefore, it is beneficial to eat such plant foods that have a high antioxidant compound content, which will cut down the incidence of certain chronic diseases, for instance diabetes, cancers and cardiovascular diseases, through the management of oxidative stress. Furthermore, berries and other fruits with low-amylase and high-glucosidase inhibitory activities could be regarded as candidate food items in the control of the early stages of hyperglycemia associated with type 2 diabetes.

Abstract: Domain Adaptation is an actively researched problem in Computer Vision. In this work, we propose an approach that leverages unsupervised data to bring the source and target distributions closer in a learned joint feature space. We accomplish this by inducing a symbiotic relationship between the learned embedding and a generative adversarial network. This is in contrast to methods which use the adversarial framework for realistic data generation and retraining deep models with such data. We demonstrate the strength and generality of our approach by performing experiments on three different tasks with varying levels of difficulty: (1) Digit classification (MNIST, SVHN and USPS datasets) (2) Object recognition using OFFICE dataset and (3) Domain adaptation from synthetic to real data. Our method achieves state-of-the art performance in most experimental settings and by far the only GAN-based method that has been shown to work well across different datasets such as OFFICE and DIGITS.

Abstract: Standard video frame interpolation methods first estimate optical flow between input frames and then synthesize an intermediate frame guided by motion. Recent approaches merge these two steps into a single convolution process by convolving input frames with spatially adaptive kernels that account for motion and re-sampling simultaneously. These methods require large kernels to handle large motion, which limits the number of pixels whose kernels can be estimated at once due to the large memory demand. To address this problem, this paper formulates frame interpolation as local separable convolution over input frames using pairs of 1D kernels. Compared to regular 2D kernels, the 1D kernels require significantly fewer parameters to be estimated. Our method develops a deep fully convolutional neural network that takes two input frames and estimates pairs of 1D kernels for all pixels simultaneously. Since our method is able to estimate kernels and synthesizes the whole video frame at once, it allows for the incorporation of perceptual loss to train the neural network to produce visually pleasing frames. This deep neural network is trained end-to-end using widely available video data without any human annotation. Both qualitative and quantitative experiments show that our method provides a practical solution to high-quality video frame interpolation.

Abstract: Plant litter represents a major basal resource in streams, where its decomposition is partly regulated by litter traits. Litter-trait variation may determine the latitudinal gradient in decomposition in streams, which is mainly microbial in the tropics and detritivore-mediated at high latitudes. However, this hypothesis remains untested, as we lack information on large-scale trait variation for riparian litter. Variation cannot easily be inferred from existing leaf-trait databases, since nutrient resorption can cause traits of litter and green leaves to diverge. Here we present the first global-scale assessment of riparian litter quality by determining latitudinal variation (spanning 107°) in litter traits (nutrient concentrations; physical and chemical defences) of 151 species from 24 regions and their relationships with environmental factors and phylogeny. We hypothesized that litter quality would increase with latitude (despite variation within regions) and traits would be correlated to produce ‘syndromes’ resulting from phylogeny and environmental variation. We found lower litter quality and higher nitrogen:phosphorus ratios in the tropics. Traits were linked but showed no phylogenetic signal, suggesting that syndromes were environmentally determined. Poorer litter quality and greater phosphorus limitation towards the equator may restrict detritivore-mediated decomposition, contributing to the predominance of microbial decomposers in tropical streams.

Abstract: Neuroinflammation is associated with neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis. Microglia and astrocytes are key regulators of inflammatory responses in the central nervous system. The activation of microglia and astrocytes is heterogeneous and traditionally categorized as neurotoxic (M1-phenotype microglia and A1-phenotype astrocytes) or neuroprotective (M2-phenotype microglia and A2-phenotype astrocytes). However, this dichotomized classification may not reflect the various phenotypes of microglia and astrocytes. The relationship between these activated glial cells is also very complicated, and the phenotypic distribution can change, based on the progression of neurodegenerative diseases. A better understanding of the roles of microglia and astrocytes in neurodegenerative diseases is essential for developing effective therapies. In this review, we discuss the roles of inflammatory response in neurodegenerative diseases, focusing on the contributions of microglia and astrocytes and their relationship. In addition, we discuss biomarkers to measure neuroinflammation and studies on therapeutic drugs that can modulate neuroinflammation.

Abstract: Three Earth-sized planets—receiving similar irradiation to Venus and Earth, and ideally suited for atmospheric study—have been found transiting a nearby ultracool dwarf star that has a mass of only eight per cent of that of the Sun. Theory predicts that terrestrial or rocky planets are likely to be orbiting the lowest-mass stars. This paper reports the detection of a system of three Earth-sized planets transiting a very nearby (12 parsec) ultracool dwarf star of only 8% of the mass of the Sun. The planets are similar in irradiation to Venus and Earth, and particularly well suited for detailed atmospheric characterization. Star-like objects with effective temperatures of less than 2,700 kelvin are referred to as ‘ultracool dwarfs’1. This heterogeneous group includes stars of extremely low mass as well as brown dwarfs (substellar objects not massive enough to sustain hydrogen fusion), and represents about 15 per cent of the population of astronomical objects near the Sun2. Core-accretion theory predicts that, given the small masses of these ultracool dwarfs, and the small sizes of their protoplanetary disks3,4, there should be a large but hitherto undetected population of terrestrial planets orbiting them5—ranging from metal-rich Mercury-sized planets6 to more hospitable volatile-rich Earth-sized planets7. Here we report observations of three short-period Earth-sized planets transiting an ultracool dwarf star only 12 parsecs away. The inner two planets receive four times and two times the irradiation of Earth, respectively, placing them close to the inner edge of the habitable zone of the star8. Our data suggest that 11 orbits remain possible for the third planet, the most likely resulting in irradiation significantly less than that received by Earth. The infrared brightness of the host star, combined with its Jupiter-like size, offers the possibility of thoroughly characterizing the components of this nearby planetary system.

Abstract: Surface nanobubbles are nanoscopic gaseous domains on immersed substrates which can survive for days. They were first speculated to exist about 20 years ago, based on stepwise features in force curves between two hydrophobic surfaces, eventually leading to the first atomic force microscopy (AFM) image in 2000. While in the early years it was suspected that they may be an artifact caused by AFM, meanwhile their existence has been confirmed with various other methods, including through direct optical observation. Their existence seems to be paradoxical, as a simple classical estimate suggests that they should dissolve in microseconds, due to the large Laplace pressure inside these nanoscopic spherical-cap-shaped objects. Moreover, their contact angle (on the gas side) is much smaller than one would expect from macroscopic counterparts. This review will not only give an overview on surface nanobubbles, but also on surface nanodroplets, which are nanoscopic droplets (e.g., of oil) on (hydrophobic) substrates immersed in water, as they show similar properties and can easily be confused with surface nanobubbles and as they are produced in a similar way, namely, by a solvent exchange process, leading to local oversaturation of the water with gas or oil, respectively, and thus to nucleation. The review starts with how surface nanobubbles and nanodroplets can be made, how they can be observed (both individually and collectively), and what their properties are. Molecular dynamic simulations and theories to account for the long lifetime of the surface nanobubbles are then reported on. The crucial element contributing to the long lifetime of surface nanobubbles and nanodroplets is pinning of the three-phase contact line at chemical or geometric surface heterogeneities. The dynamical evolution of the surface nanobubbles then follows from the diffusion equation, Laplace’s equation, and Henry’s law. In particular, one obtains stable surface nanobubbles when the gas influx from the gas-oversaturated water and the outflux due to Laplace pressure balance. This is only possible for small enough surface bubbles. It is therefore the gas or oil oversaturation ζ that determines the contact angle of the surface nanobubble or nanodroplet and not the Young equation. The review also covers the potential technological relevance of surface nanobubbles and nanodroplets, namely, in flotation, in (photo)catalysis and electrolysis, in nanomaterial engineering, for transport in and out of nanofluidic devices, and for plasmonic bubbles, vapor nanobubbles, and energy conversion. Also given is a discussion on surface nanobubbles and nanodroplets in a nutshell, including theoretical predictions resulting from it and future directions. Studying the nucleation, growth, and dissolution dynamics of surface nanobubbles and nanodroplets will shed new light on the problems of contact line pinning and contact angle hysteresis on the submicron scale.

Abstract: The seasonal cycle of respiratory viral diseases has been widely recognized for thousands of years, as annual epidemics of the common cold and influenza disease hit the human population like clockwork in the winter season in temperate regions. Moreover, epidemics caused by viruses such as severe acute respiratory syndrome coronavirus (SARS-CoV) and the newly emerging SARS-CoV-2 occur during the winter months. The mechanisms underlying the seasonal nature of respiratory viral infections have been examined and debated for many years. The two major contributing factors are the changes in environmental parameters and human behavior. Studies have revealed the effect of temperature and humidity on respiratory virus stability and transmission rates. More recent research highlights the importance of the environmental factors, especially temperature and humidity, in modulating host intrinsic, innate, and adaptive immune responses to viral infections in the respiratory tract. Here we review evidence of how outdoor and indoor climates are linked to the seasonality of viral respiratory infections. We further discuss determinants of host response in the seasonality of respiratory viruses by highlighting recent studies in the field. Expected final online publication date for the Annual Review of Virology, Volume 7 is September 29, 2020. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Abstract: We present a growing collection of Android Applications col-lected from several sources, including the official GooglePlay app market. Our dataset, AndroZoo, currently contains more than three million apps, each of which has beenanalysed by tens of different AntiVirus products to knowwhich applications are detected as Malware. We provide thisdataset to contribute to ongoing research efforts, as well asto enable new potential research topics on Android Apps.By releasing our dataset to the research community, we alsoaim at encouraging our fellow researchers to engage in reproducible experiments.

Abstract: This review covers the recent literature on the geography of sustainability transitions and takes stock with achieved theoretical and empirical insights. The review synthesises and reflects upon insights of relevance for sustainability transitions following from analyses of the importance of place specificity and the geography of inter-organisational relations. It is found that these contributions focus on the geography of niche development rather than regime dynamics, and that there is an emphasis on understanding the importance of place-specificity at the local level. While there is a wide consensus that place-specificity matters there is still little generalisable knowledge about how place-specificity matters for transitions. Most contributions add spatial sensitivity to frameworks from the transitions literature, but few studies suggest alternative frameworks to study sustainability transitions. To address this, the review suggests promising avenues for future research on the geography of sustainability transitions, drawing on recent theoretical advancements in economic geography.

Abstract: Since their inception in 2011, from the inaugural synthesis of multi-layered Ti3C2Tx by etching Ti3AlC2 with hydrofluoric acid (HF), novel routes with a myriad of reducing agents, etchants and intercalants have been explored and many new members have been added to the two-dimensional (2D) material constellation. Despite being endowed with the rare combination of good electronic conductivity and hydrophilicity, their inherent low capacities, for instance, temper their prospect for application for electrodes in energy storage systems. MXene-based composites, however, with a probable synergistic effect in agglomeration prevention, facilitating electronic conductivity, improving electrochemical stability, enhancing pseudocapacitance and minimizing the shortcomings of individual components, exceed the previously mentioned capacitance ceiling. In this review, we summarise the development and progress in the synthesis of various multi-layered carbides, carbonitrides and nitrides, and intercalants, as well as the subsequent processing in order to delaminate them into single- and/or few-layered and incorporate into MXene-based composites, focusing on their performance and application as transparent conductive films, environmental remediation, electromagnetic interference (EMI) absorption and shielding, electrocatalysts, Li-ion batteries (LIBs), supercapacitors and other electrochemical storage systems.

Abstract: Some pathogens and pests deliver small RNAs (sRNAs) into host cells to suppress host immunity. Conversely, hosts also transfer sRNAs into pathogens and pests to inhibit their virulence. Although sRNA trafficking has been observed in a wide variety of interactions, how sRNAs are transferred, especially from hosts to pathogens and pests, is still unknown. Here, we show that host Arabidopsis cells secrete exosome-like extracellular vesicles to deliver sRNAs into fungal pathogen Botrytis cinerea . These sRNA-containing vesicles accumulate at the infection sites and are taken up by the fungal cells. Transferred host sRNAs induce silencing of fungal genes critical for pathogenicity. Thus, Arabidopsis has adapted exosome-mediated cross-kingdom RNA interference as part of its immune responses during the evolutionary arms race with the pathogen.

Abstract: The transition between soluble intrinsically disordered tau protein and aggregated tau in neurofibrillary tangles in Alzheimer's disease is unknown. Here, we propose that soluble tau species can undergo liquid–liquid phase separation (LLPS) under cellular conditions and that phase‐separated tau droplets can serve as an intermediate toward tau aggregate formation. We demonstrate that phosphorylated or mutant aggregation prone recombinant tau undergoes LLPS, as does high molecular weight soluble phospho‐tau isolated from human Alzheimer brain. Droplet‐like tau can also be observed in neurons and other cells. We found that tau droplets become gel‐like in minutes, and over days start to spontaneously form thioflavin‐S‐positive tau aggregates that are competent of seeding cellular tau aggregation. Since analogous LLPS observations have been made for FUS, hnRNPA1, and TDP43, which aggregate in the context of amyotrophic lateral sclerosis, we suggest that LLPS represents a biophysical process with a role in multiple different neurodegenerative diseases.

Abstract: In [1] we gave a precise holographic calculation of chaos at the scrambling time scale. We studied the influence of a small perturbation, long in the past, on a two-sided correlation function in the thermofield double state. A similar analysis applies to squared commutators and other out-of-time-order one-sided correlators [2-6]. The essential bulk physics is a high energy scattering problem near the horizon of an AdS black hole. The above papers used Einstein gravity to study this problem; in the present paper we consider stringy and Planckian corrections. Elastic stringy corrections play an important role, effectively weakening and smearing out the development of chaos. We discuss their signature in the boundary field theory, commenting on the extension to weak coupling. Inelastic effects, although important for the evolution of the state, leave a parametrically small imprint on the correlators that we study. We briefly discuss ways to diagnose these small corrections, and we propose another correlator where inelastic effects are order one.