A new unsupervised convolutional neural network model for Chinese scene text detection

As one of the most popular deep learning models, convolution neural network (CNN) has achieved huge success in image information extraction. Traditionally CNN is trained by supervised learning method with labeled data and used as a classifier by adding a classification layer in the end. Its capability of extracting image features is largely limited due to the difficulty of setting up a large training dataset. In this paper, we propose a new unsupervised learning CNN model, which uses a so-called convolutional sparse auto-encoder (CSAE) algorithm pre-Train the CNN. Instead of using labeled natural images for CNN training, the CSAE algorithm can be used to train the CNN with unlabeled artificial images, which enables easy expansion of training data and unsupervised learning. The CSAE algorithm is especially designed for extracting complex features from specific objects such as Chinese characters. After the features of articficial images are extracted by the CSAE algorithm, the learned parameters are used to initialize the first CNN convolutional layer, and then the CNN model is fine-Trained by scene image patches with a linear classifier. The new CNN model is applied to Chinese scene text detection and is evaluated with a multilingual image dataset, which labels Chinese, English and numerals texts separately. More than 10% detection precision gain is observed over two CNN models.

A novel neural network technique for exploratory data analysis

This paper surveys the context of feature extraction by neural network approaches, and compares and contrasts their behaviour as prospective data visualisation tools in a real world problem. We also introduce and discuss a hybrid approach which allows us to control the degree of discriminatory and topographic information in the extracted feature space.

Learning task specific distributed paragraph representations using a 2-tier convolutional neural network

We introduce a type of 2-tier convolutional neural network model for learning distributed paragraph representations for a special task (e.g. paragraph or short document level sentiment analysis and text topic categorization). We decompose the paragraph semantics into 3 cascaded constitutes: word representation, sentence composition and document composition. Specifically, we learn distributed word representations by a continuous bag-of-words model from a large unstructured text corpus. Then, using these word representations as pre-trained vectors, distributed task specific sentence representations are learned from a sentence level corpus with task-specific labels by the first tier of our model. Using these sentence representations as distributed paragraph representation vectors, distributed paragraph representations are learned from a paragraph-level corpus by the second tier of our model. It is evaluated on DBpedia ontology classification dataset and Amazon review dataset. Empirical results show the effectiveness of our proposed learning model for generating distributed paragraph representations.

Design and analysis of distributed utility maximization algorithm for multihop wireless network with inaccurate feedback

Distributed network utility maximization (NUM) is receiving increasing interests for cross-layer optimization problems in multihop wireless networks. Traditional distributed NUM algorithms rely heavily on feedback information between different network elements, such as traffic sources and routers. Because of the distinct features of multihop wireless networks such as time-varying channels and dynamic network topology, the feedback information is usually inaccurate, which represents as a major obstacle for distributed NUM application to wireless networks. The questions to be answered include if distributed NUM algorithm can converge with inaccurate feedback and how to design effective distributed NUM algorithm for wireless networks. In this paper, we first use the infinitesimal perturbation analysis technique to provide an unbiased gradient estimation on the aggregate rate of traffic sources at the routers based on locally available information. On the basis of that, we propose a stochastic approximation algorithm to solve the distributed NUM problem with inaccurate feedback. We then prove that the proposed algorithm can converge to the optimum solution of distributed NUM with perfect feedback under certain conditions. The proposed algorithm is applied to the joint rate and media access control problem for wireless networks. Numerical results demonstrate the convergence of the proposed algorithm. © 2013 John Wiley & Sons, Ltd.

Gene network and proteomic analyses of cardiac responses to pathological and physiological stress

Background—The molecular mechanisms underlying similarities and differences between physiological and pathological left ventricular hypertrophy (LVH) are of intense interest. Most previous work involved targeted analysis of individual signaling pathways or screening of transcriptomic profiles. We developed a network biology approach using genomic and proteomic data to study the molecular patterns that distinguish pathological and physiological LVH. Methods and Results—A network-based analysis using graph theory methods was undertaken on 127 genome-wide expression arrays of in vivo murine LVH. This revealed phenotype-specific pathological and physiological gene coexpression networks. Despite >1650 common genes in the 2 networks, network structure is significantly different. This is largely because of rewiring of genes that are differentially coexpressed in the 2 networks; this novel concept of differential wiring was further validated experimentally. Functional analysis of the rewired network revealed several distinct cellular pathways and gene sets. Deeper exploration was undertaken by targeted proteomic analysis of mitochondrial, myofilament, and extracellular subproteomes in pathological LVH. A notable finding was that mRNA–protein correlation was greater at the cellular pathway level than for individual loci. Conclusions—This first combined gene network and proteomic analysis of LVH reveals novel insights into the integrated pathomechanisms that distinguish pathological versus physiological phenotypes. In particular, we identify differential gene wiring as a major distinguishing feature of these phenotypes. This approach provides a platform for the investigation of potentially novel pathways in LVH and offers a freely accessible protocol (http://sites.google.com/site/cardionetworks) for similar analyses in other cardiovascular diseases.

Network learning:Exploring learning by interorganizational networks

The importance of interorganizational networks in supporting or hindering the achievement of organizational objectives is now widely acknowledged. Network research is directed at understanding network processes and structures, and their impact upon performance. A key process is learning. The concepts of individual, group and organizational learning are long established. This article argues that learning might also usefully be regarded as occurring at a fourth system level, the interorganizational network. The concept of network learning - learning by a group of organizations as a group - is presented, and differentiated from other types of learning, notably interorganizational learning (learning in interorganizational contexts). Four cases of network learning are identified and analysed to provide insights into network learning processes and outcomes. It is proposed that 'network learning episode' offers a suitable unit of analysis for the empirical research needed to develop our understanding of this potentially important concept.

Design and analysis of distributed utility maximization algorithm for multihop wireless network with inaccurate feedback

Distributed network utility maximization (NUM) is receiving increasing interests for cross-layer optimization problems in multihop wireless networks. Traditional distributed NUM algorithms rely heavily on feedback information between different network elements, such as traffic sources and routers. Because of the distinct features of multihop wireless networks such as time-varying channels and dynamic network topology, the feedback information is usually inaccurate, which represents as a major obstacle for distributed NUM application to wireless networks. The questions to be answered include if distributed NUM algorithm can converge with inaccurate feedback and how to design effective distributed NUM algorithm for wireless networks. In this paper, we first use the infinitesimal perturbation analysis technique to provide an unbiased gradient estimation on the aggregate rate of traffic sources at the routers based on locally available information. On the basis of that, we propose a stochastic approximation algorithm to solve the distributed NUM problem with inaccurate feedback. We then prove that the proposed algorithm can converge to the optimum solution of distributed NUM with perfect feedback under certain conditions. The proposed algorithm is applied to the joint rate and media access control problem for wireless networks. Numerical results demonstrate the convergence of the proposed algorithm. © 2013 John Wiley & Sons, Ltd.

Numerical analysis of artificial neural network and volterra-based nonlinear equalizers for coherent optical OFDM

One major drawback of coherent optical orthogonal frequency-division multiplexing (CO-OFDM) that hitherto remains unsolved is its vulnerability to nonlinear fiber effects due to its high peak-to-average power ratio. Several digital signal processing techniques have been investigated for the compensation of fiber nonlinearities, e.g., digital back-propagation, nonlinear pre- and post-compensation and nonlinear equalizers (NLEs) based on the inverse Volterra-series transfer function (IVSTF). Alternatively, nonlinearities can be mitigated using nonlinear decision classifiers such as artificial neural networks (ANNs) based on a multilayer perceptron. In this paper, ANN-NLE is presented for a 16QAM CO-OFDM system. The capability of the proposed approach to compensate the fiber nonlinearities is numerically demonstrated for up to 100-Gb/s and over 1000km and compared to the benchmark IVSTF-NLE. Results show that in terms of Q-factor, for 100-Gb/s at 1000km of transmission, ANN-NLE outperforms linear equalization and IVSTF-NLE by 3.2dB and 1dB, respectively.

Managing supply chain knowledge in the new product development process:a social network analysis approach

The aim of this paper is to propose a conceptual framework for studying the knowledge transfer problem within the supply chain. The social network analysis (SNA) is presented as a useful tool to study knowledge networks within supply chain, to visualize knowledge flows and to identify the accumulating knowledge nodes of the networks. © 2011 IEEE.

Understanding inter-firm networks and types of innovation in SMEs:a social network perspective

Innovation is one of the key drivers for gaining competitive advantages in any firms. Understanding knowledge transfer through inter-firm networks and its effects on types of innovation in SMEs is very important in improving SMEs innovation. This study examines relationships between characteristics of inter-firm knowledge transfer networks and types of innovation in SMEs. To achieve this, social network perspective is adopted to understand inter-firm knowledge transfer networks and its impact on innovation by investigating how and to what extend ego network characteristics are affecting types of innovation. Therefore, managers can develop the firms'network according to their strategies and requirements. First, a conceptual model and research hypotheses are proposed to establish the possible relationship between network properties and types of innovation. Three aspects of ego network are identified and adopted for hypotheses development: 1) structural properties which address the potential for resources and the context for the flow of resources, 2) relational properties which reflect the quality of resource flows, and 3) nodal properties which are about quality and variety of resources and capabilities of the ego partners. A questionnaire has been designed based on the hypotheses. Second, semistructured interviews with managers of five SMEs have been carried out, and a thematic qualitative analysis of these interviews has been performed. The interviews helped to revise the questionnaire and provided preliminary evidence to support the hypotheses. Insights from the preliminary investigation also helped to develop research plan for the next stage of this research.

Performance analysis of an experimental wireless relay sensor network

Communication through relay channels in wireless sensor networks can create diversity and consequently improve the robustness of data transmission for ubiquitous computing and networking applications. In this paper, we investigate the performances of relay channels in terms of diversity gain and throughput via both experimental research and theoretical analysis. Two relaying algorithms, dynamic relaying and fixed relaying, are utilised and tested to find out what the relay channels can contribute to system performances. The tests are based on a wireless relay sensor network comprising a source node, a destination node and a couple of relay nodes, and carried out in an indoor environment with rare movement of objects nearby. The tests confirm, in line with the analytical results, that more relay nodes lead to higher diversity gain in the network. The test results also show that the data throughput between the source node and the destination node is enhanced by the presence of the relay nodes. Energy consumption in association with the relaying strategy is also analysed. Copyright © 2009 John Wiley & Sons, Ltd.