Search

To improve the ability of natural language processing (NLP) in mobile communication networks, a mobile communication network NLP system based on a recurrent neural network is designed in the eleventh paper, "A Novel Natural Language Processing Model in Mobile Communication Networks", authored by Gautam Srivastava from the Department of Mathematics and Computer Science, Brandon University, Canada. First of all, image and visual processing in mobile networks are the key technology in the research of information processing because 90% of human information got from world comes from vision [[1]]. Then, the varied applications of information processing under mobile network environment are other key topics in research because the application ecology determines the human acceptability of the digital world [[5]]. It is a kind of huge virtual digital world to people that the theories and applications of information processing have been changed to complex multimodal information processing. [Extracted from the article]

Deploying convolutional neural networks (CNNs) on mobile devices is difficult due to the limited memory and computation resources. We aim to design efficient neural networks for heterogeneous devices including CPU and GPU, by exploiting the redundancy in feature maps, which has rarely been investigated in neural architecture design. For CPU-like devices, we propose a novel CPU-efficient Ghost (C-Ghost) module to generate more feature maps from cheap operations. Based on a set of intrinsic feature maps, we apply a series of linear transformations with cheap cost to generate many ghost feature maps that could fully reveal information underlying intrinsic features. The proposed C-Ghost module can be taken as a plug-and-play component to upgrade existing convolutional neural networks. C-Ghost bottlenecks are designed to stack C-Ghost modules, and then the lightweight C-GhostNet can be easily established. We further consider the efficient networks for GPU devices. Without involving too many GPU-inefficient operations (e.g., depth-wise convolution) in a building stage, we propose to utilize the stage-wise feature redundancy to formulate GPU-efficient Ghost (G-Ghost) stage structure. The features in a stage are split into two parts where the first part is processed using the original block with fewer output channels for generating intrinsic features, and the other are generated using cheap operations by exploiting stage-wise redundancy. Experiments conducted on benchmarks demonstrate the effectiveness of the proposed C-Ghost module and the G-Ghost stage. C-GhostNet and G-GhostNet can achieve the optimal trade-off of accuracy and latency for CPU and GPU, respectively. MindSpore code is available at https://gitee.com/mindspore/models/pulls/1809, and PyTorch code is available at https://github.com/huawei-noah/CV-Backbones. [ABSTRACT FROM AUTHOR]

High dimensionality and sample imbalance of gene expression data promote the development of effective algorithms for classifying gene expression data. To improve the ability to distinguish different subtypes of gene expression data, we devise a hypervolume-based discrete evolutionary optimization algorithm (HYBDEOA) in this paper. Four objectives, namely the number of genes, the accuracy, the relevance, and the redundancy, are optimized simultaneously to guide the evolution. Firstly, binary encoding is used to choose some features, projecting data onto different subspaces. After that, a discrete neighborhood operation is conducted to generate a new binary-mapped population. Combining the new population with the current population, we employ the hypervolume-based mechanism to select the Pareto solutions. Finally, a discrete mutation method is proposed to find promising solutions in the binary search space. To demonstrate the performance of HYBDEOA, we apply HYBDEOA to 55 synthetic datasets and 35 cancer gene expression datasets. Extensive experiments are also conducted to reveal the effectiveness and efficiency of HYBDEOA. The experimental results demonstrate that our proposed method is a parameter-less and robust algorithm, which can group gene expression data with a finer and more informative classification. [ABSTRACT FROM AUTHOR]

Arctic sea ice cover has decreased dramatically over the last three decades. This study quantifies the sea ice concentration (SIC) trends in the Arctic Ocean over the period of 1979–2016 and analyzes their spatial and temporal variations. During each month the SIC trends are negative over the Arctic Ocean, wherein the largest (smallest) rate of decline found in September (March) is -0.48%/a (-0.10%/a). The summer (-0.42%/a) and autumn (-0.31%/a) seasons show faster decrease rates than those of winter (-0.12%/a) and spring (-0.20%/a) seasons. Regional variability is large in the annual SIC trend. The largest SIC trends are observed for the Kara (-0.60%/a) and Barents Seas (-0.54%/a), followed by the Chukchi Sea (-0.48%/a), East Siberian Sea (-0.43%/a), Laptev Sea (-0.38%/a), and Beaufort Sea (-0.36%/a). The annual SIC trend for the whole Arctic Ocean is -0.26%/a over the same period. Furthermore, the infl uences and feedbacks between the SIC and three climate indexes and three climatic parameters, including the Arctic Oscillation (AO), North Atlantic Oscillation (NAO), Dipole anomaly (DA), sea surface temperature (SST), surface air temperature (SAT), and surface wind (SW), are investigated. Statistically, sea ice provides memory for the Arctic climate system so that changes in SIC driven by the climate indices (AO, NAO and DA) can be felt during the ensuing seasons. Positive SST trends can cause greater SIC reductions, which is observed in the Greenland and Barents Seas during the autumn and winter. In contrast, the removal of sea ice (i.e., loss of the insulating layer) likely contributes to a colder sea surface (i.e., decreased SST), as is observed in northern Barents Sea. Decreasing SIC trends can lead to an in-phase enhancement of SAT, while SAT variations seem to have a lagged infl uence on SIC trends. SW plays an important role in the modulating SIC trends in two ways: by transporting moist and warm air that melts sea ice in peripheral seas (typically evident inthe Barents Sea) and by exporting sea ice out of the Arctic Ocean via passages into the Greenland and Barents Seas, including the Fram Strait, the passage between Svalbard and Franz Josef Land (S-FJL), and the passage between Franz Josef Land and Severnaya Zemlya (FJL-SZ). [ABSTRACT FROM AUTHOR]

The transition set semantics (Wang and Jiao, LNCS 6128:84-103, ) partitions the Petri net behaviors in a canonical way such that behaviors in an equivalence class have the same canonical transition set sequence. This article extends the semantics in two ways: firstly, the semantics is parameterized by the basic relation on the structural transitions to define different variants; secondly, the semantics for the infinite firing sequences of the net is defined. We prove that these extensions still preserve the well-definedness, soundness and completeness of the semantics. Furthermore, we show how to recognize some infinite sequences called back-loops in the view of this new semantics. [ABSTRACT FROM AUTHOR]

In comparison with seasonal sea ice (first-year ice, FY ice), multiyear (MY) sea ice is thicker and has more opportunity to survive through the summer melting seasons. Therefore, the variability of wintertime MY ice plays a vital role in modulating the variations in the Arctic sea ice minimum extent during the following summer. As a response, the ice-ocean-atmosphere interactions may be significantly affected by the variations in the MY ice cover. Satellite observations are characterized by their capability to capture the spatiotemporal changes of Arctic sea ice. During the recent decades, many active and passive sensors onboard a variety of satellites (QuikSCAT, ASCAT, SSMIS, ICESat, CryoSat-2, etc.) have been used to monitor the dramatic loss of Arctic MY ice. The main objective of this study is to outline the advances and remaining challenges in monitoring the MY ice changes through the utilization of multiple satellite observations. We summarize the primary satellite data sources that are used to identify MY ice. The methodology to classify MY ice and derive MY ice concentration is reviewed. The interannual variability and trends in the MY ice time series in terms of coverage, thickness, volume, and age composition are evaluated. The potential causes associated with the observed Arctic MY ice loss are outlined, which are primarily related to the export and melting mechanisms. In addition, the causes to the MY ice depletion from the perspective of the oceanic water inflow from Pacific and Atlantic Oceans and the water vapor intrusion, as well as the roles of synoptic weather, are analyzed. The remaining challenges and possible upcoming research subjects in detecting the rapidly changing Arctic MY ice using the combined application of multisource remote sensing techniques are discussed. Moreover, some suggestions for the future application of satellite observations on the investigations of MY ice cover changes are proposed. [ABSTRACT FROM AUTHOR]

Key message: A novel genomic region controlling thermotolerance at flowering was identified by the combination of whole genomic re-sequencing and bulked segregant analysis in maize. The increasing frequency of extreme high temperature has brought a great threat to the development of maize throughout its life cycle, especially during the flowering phase. However, the genetic basis of thermotolerance at flowering in maize remains poorly understood. Here, we characterized a thermotolerant maize ecotype Abe2 and dissected its genetic basis using a F2:8 recombinant inbred line (RIL) population generated from a cross between Abe2 and B73. After continuous high temperature stress above 35 °C for 17 days, Abe2 and B73 show distinct leaf scorching phenotype under field conditions. To identify the genomic regions associated with the phenotypic variation, we applied a combination of whole genomic re-sequencing and bulked segregant analysis, and revealed 10,316,744 SNPs and 1,488,302 InDels between the two parental lines, and 2,693,054 SNPs and 313,757 InDels between the two DNA pools generated from the thermos-tolerant and the sensitive individuals of the RIL, of which, 108,655 and 17,853 SNPs may cause nonsynonymous variations. Finally, a 7.41 Mb genomic region on chromosome 1 was identified, and 7 candidate genes were annotated to participate in high temperature-related stress response. A candidate gene Zm00001d033339 encoding a serine/threonine protein kinase was proposed to be the most likely causative gene contributing to the thermotolerance at flowering by involving in stomatal movement (GO: 0010119) via Abscisic acid (ABA) pathway (KO04075). This work could provide an opportunity for gene cloning and pyramiding breeding to improve thermotolerance at flowering in maize. [ABSTRACT FROM AUTHOR]