Search

Single-inductor multiple-output (SIMO) voltage regulators allow multiple voltage domains to share a single inductor, thus representing a domain-scalable approach to energy-efficient integrated voltage regulation (IVR). However, poor transient response and significant supply voltage ( $V_{dd}$ ) ripple in SIMO regulators induce severe voltage margins. This article quantifies the prohibitive energy-efficiency impact of these margins and proposes two techniques to address them: dynamic droop allocation (DDA) through concurrent domain- $V_{dd}$ control and adaptive clocking using the UniCap architecture. We demonstrate the effectiveness of both techniques on an integrated four-domain SIMO system on chip (SoC) in 65-nm CMOS. Measurements indicate that, compared to conventional SIMO implementations, $V_{dd}$ guardband reductions obtained by UniCap (98%) and DDA (40%) reduce total system power draw by 53% and 31% respectively.

Neural Sign Language Translation (SLT), which is an important cross-modal task to bridge the communication gap between the deaf and the hearing people, has attracted great attention in the field of artificial intelligence, computer vision, and multimedia, etc. Although some great progress has been achieved recently, current neural SLT models still suffer from translation errors caused by under-consideration of non-manual features such as facial expressions, which can carry critical information during communication among the deaf. This paper aims to enhance the traditional neural SLT models by highlighting the facial expression information in the CNN-based sign video representing part. Two novel schemes have been proposed. The first scheme is based on a Multi-stream Architecture, which extracts and represents the facial expression information in an additional stream and aggregates it with the information from the main stream. The second scheme is a pre-trained scheme based on Regions of Interest (RoIs), which first trains a multi-region detection module for recognizing the features of faces and bodies and then transfers the pre-trained parameters to the module in SLT model. To validate the proposed models, we conducted the experiments upon the publicly available SLT benchmark dataset: RWTH-PHOENIX-Weather-2014T. Experimental results showed that both the above-mentioned schemes can improve the performance of SLT models. Especially, the RoIs-based scheme can achieve an improvement up to 1.6+ BLEU-4 score gains, while the multi-stream scheme quantitatively analyzed the importance of the face mainly through flexible components, providing a sufficient theoretical basis for the RoIs-based scheme.

Knowledge representation is one of the critical problems in knowledge engineering and artificial intelligence, while knowledge embedding as a knowledge representation methodology indicates entities and relations in knowledge graph as low-dimensional, continuous vectors. In this way, knowledge graph is compatible with numerical machine learning models. Major knowledge embedding methods employ geometric translation to design score function, which is weak-semantic for natural language processing. To overcome this disadvantage, in this paper, we propose our model based on multi-view clustering framework, which could generate semantic representations of knowledge elements (i.e., entities/relations ). With our semantic model, we also present an empowered solution to entity retrieval with entity description. Extensive experiments show that our model achieves substantial improvements against baselines on the task of knowledge graph completion, triple classification, entity classification, and entity retrieval.

With the liberalization of incremental distribution network investment business, more and more investment projects from various sources and diversified ownership are eligible for construction applications. This article proposes a many-criteria investment priorities evaluation method for incremental distribution network planning based on the hyperplane projection transformation. In the established many-criteria evaluation indicator system, the multi-grade indicators including economic and technical factors are adopted to comprehensively evaluate the distribution network infrastructure projects. Furthermore, the hyperplane projection transformation is applied to project the solutions from the many-objective hypercube to the normalized hyperplane. Then, the comprehensive distance quantifies the quality and balance of the distribution network infrastructure project indicators, and the investment priority rankings can be obtained. Finally, the performance of the proposed method is validated on six distribution network infrastructure projects with eleven evaluation indicators. The results show that the proposed many-criteria evaluation of infrastructure investment priorities is effective to recognize the non-dominance solutions with extremely good or extremely poor indicators.

Implicit discourse relation recognition is the performance bottleneck of discourse structure analysis. To alleviate the shortage of training data, previous methods usually use explicit discourse data, which are naturally labeled by connectives, as additional training data. However, it is often difficult for them to integrate large amounts of explicit discourse data because of the noise problem. In this paper, we propose a simple and effective method to leverage massive explicit discourse data. Specifically, we learn connective-based word embeddings (CBWE) by performing connective classification on explicit discourse data. The learned CBWE is capable of capturing discourse relationships between words, and can be used as pre-trained word embeddings for implicit discourse relation recognition. On both the English PDTB and Chinese CDTB data sets, using CBWE achieves significant improvements over baselines with general word embeddings, and better performance than baselines integrating explicit discourse data. By combining CBWE with a strong baseline, we achieve the state-of-the-art performance.

In this paper, an uncertainty and disturbance estimator (UDE)-based position controller for quadrotors in the absence of velocity measurements is presented. Firstly, the nonlinear dynamics of a quadrotor is derived and simplified into a decoupled control design model using small angle approximation. Secondly, an UDE is constructed to compensate the system uncertainties and external disturbances. By applying passivity techniques to produce damping in the system, the proposed UDE-based controller can achieve trajectory tracking without velocity measurements and has a good robust performance. Moreover, the stability analysis of the closed-loop system is also introduced. Finally, the proposed method and the control method based on classical uncertainty and disturbance estimator are implemented and compared in the Gazebo simulation and flight experiment with a real quadrotor platform, are implemented and compared in the Gazebo simulation and flight experiment with a real quadrotor platform. Both simulation and experiment results demonstrate the effectiveness and advantages of the proposed control method in signal tracking and disturbance rejection.

Although neural machine translation (NMT) has certain capability to implicitly learn semantic information of sentences, we explore and show that Part-of-Speech (POS) tags can be explicitly incorporated into the attention mechanism of NMT effectively to yield further improvements. In this article, we propose an NMT model with tag-enhanced attention mechanism. In our model, NMT and POS tagging are jointly modeled via multi-task learning. Besides following common practice to enrich encoder annotations by introducing predicted source POS tags, we exploit predicted target POS tags to refine attention model in a coarse-to-fine manner. Specifically, we first implement a coarse attention operation solely on source annotations and target hidden state, where the produced context vector is applied to update target hidden state used for target POS tagging. Then, we perform a fine attention operation that extends the coarse one by further exploiting the predicted target POS tags. Finally, we facilitate word prediction by simultaneously utilizing the context vector from fine attention and the predicted target POS tags. Experimental results and further analyses on Chinese-English and Japanese-English translation tasks demonstrate the superiority of our proposed model over the conventional NMT models. We release our code at https://github.com/middlekisser/PEA-NMT.git.

As with the huge number of deaf-mute people in China is of concern, there is a growing need to integrate them into mainstream society through the use of efficient sign language processing technologies. Sign language processing entails the systematic recognition and translation of sign language images/videos to text or speech. This survey provides an overview of the most important work on Chinese sign language recognition and translation, discussed its classification, highlights the features explored in sign language recognition research, presents the datasets available, and provides trends for the future research.

With the rapid development of single-cell sequencing techniques, several large-scale cell atlas projects have been launched across the world. However, it is still challenging to integrate single-cell RNA-seq (scRNA-seq) datasets with diverse tissue sources, developmental stages and/or few overlaps, due to the ambiguity in determining the batch information, which is particularly important for current batch-effect correction methods. Here, we present SCORE, a simple network-based integration methodology, which incorporates curated molecular network features to infer cellular states and generate a unified workflow for integrating scRNA-seq datasets. Validating on real single-cell datasets, we showed that regardless of batch information, SCORE outperforms existing methods in accuracy, robustness, scalability and data integration.

Due to the limited road resources and the ever-increasing number of vehicles and persons, more frequent traffic violations and higher management costs have resulted. It is crucial to propose a more intelligent and less cost scheme to solve the traffic management problem. In this paper, we design and implement a vehicle violation detection system based on deep learning, which includes the detection, tracking and recognition of vehicles. On this basis, the detection and real-time alarm of common violations, such as red light running and impolite pedestrian, are also supported. Compared with the traditional detection and monitoring based on physical equipment, our system is completely based on computer vision, where the cutting-edge achievements of deep learning have been improved and applied. The system is not only more intelligent, but also can reduce the cost to a greater extent. Experiments illustrate that the system can meet the needs of the intelligent management of urban traffic through real-time monitoring and data analysis of the traffic scenes.

Back-translation has been proven to be an effective data augmentation method that translates target monolingual data into source-side to improve the performance of Neural Machine Translation (NMT), especially in low-resource scenarios. Previous researches show that diversity of the synthetic source sentences is essential for back-translation. However, the frequently used random methods such as sampling or noised beam search, although can output diverse back-translations, often generate noisy synthetic sentences. To alleviate this problem, we propose a simple but effective constraint random decoding method for back-translation. The proposed method is based on an automatic post-editing (APE) data augment framework, which incorporates fluency boost learning. Moreover, to increase the diversity of synthetic data and ensure quality, we proposed to use an evolution decoding algorithm. Compared with the original back-translation, our method can generate more diverse while less noisy synthetic sentences. The experimental results show that the proposed method can get 0.6 BLEU improvements on the WMT18 EN-DE news dataset and more than 0.4 BLEU improvements on the EN-ZH dataset which is in the medical field, respectively.

Machine translation Quality Estimation (QE) aims to estimate the quality of machine translations without relying on golden references. Current QE researches mainly focus on sentence-level QE models, which could not capture discourse-related translation errors. To tackle this problem, this paper presents a novel document-level QE model based on Centering Theory (CT), which is a linguistics theory for assessing discourse coherence. Furthermore, we construct and release an open-source Chinese-English corpus at https://github.com/ydc/cpqe for document-level machine translation QE, which could be used to support further studies. Finally, experimental results show that the proposed model significantly outperformed the baseline model.

User reviews contain rich information about user interests in items. Recently, many deep learning methods have attempted to integrate review contents into user preference rating prediction, helping to solve data sparseness problems. However, existing methods suffered from an inherent limitation that many reviews are noisy and contain non-consecutive viewpoints, besides, they are insufficient to capture inter-viewpoint relationships. Incorporating useful information is helpful for more accurate recommendations. In this paper, we propose a neural recommendation approach with a Diversity Penalty mechanism and Capsule Networks, named DPCN. Specifically, the diversity penalty component employs weight distributed matrices with the penalization term to capture different viewpoints in textual reviews. The capsule networks are designed to aggregate individual viewpoint vectors to form high-level feature representations for feature interaction. Then we combine the review feature representations with the user and item ID embedding for final rating prediction. Extensive experiments on five real-world datasets validate the effectiveness of our approach.

Discourse topic structure is the key to the cohesion of the discourse and reflects the essence of the text. Current Chinese discourse corpus are constructed mainly based on rhetoric and semantic relations, which ignore the functional information in discourse. To alleviate this problem, we introduce a new Chinese discourse analysis dataset called CTRD, which stands for Chinese Theme-Rheme Discourse dataset. Different from previous discourse banks, CTRD was annotated according to a novel discourse annotation scheme based on the Chinese theme-rheme theory and thematic progression patterns from Halliday’s systemic functional grammar. As a result, we manually annotated 525 news documents from OntoNotes 4.0 with a Kappa value greater than 0.6. And preliminary experiments on this corpus verify the computability of CTRD. Finally, we make CTRD available at https://github.com/ydc/ctrd.

This paper describes our two systems submitted to the simultaneous translation evaluation at the 2nd automatic simultaneous translation workshop.

Aiming at the high complexity of parameter optimization for portfolio models, this paper designs a distributed high-performance portfolio optimization platform(HPPO) based on parallel computing framework and event driven architecture. The platform consists of the data layer, the model layer, and the excursion layer, which is built in a component, pluggable, and loosely coupled way. The platform adopts parallelization acceleration for backtesting and optimizing parameters of portfolio models in a certain historical interval. The platform is able to docking portfolio model with real-time market. Based on the HPPO platform, a parallel program is designed to optimize the parameters of the value at risk(VAR) model. The performance of the platform are summarized by analyzing the experimental results and comparing with the open source framework Zipline and Rqalpha.

At present, the research on target 3D reconstruction and pose measurement is hot issue in the field of computer vision. This paper proposes a method of target 3D reconstruction and pose measurement. It is considered that the 3D information of scene can be recovered by using the visible binocular camera according to the geometric relationship of corresponding points between images. The amount of information is large and the details are rich. However, due to the image as the input data, it brings sensitivity to the change of environmental light Sense, the weak texture region is difficult to match. Lidar can directly measure the three-dimensional information of the scene by emitting laser, which has the advantages of accurate measurement and little influence by external environment changes, but it has the disadvantages of sparse data and low scanning frequency. Based on the characteristics of binocular camera and lidar, aiming at the 3D reconstruction and pose measurement of the target, this paper studies the 3D structure restoration method of the spatial target based on lidar and visible light, and studies the pose measurement method by identifying the typical features of the target, so as to provide a solution for the 3D reconstruction and measurement of the object.

Background Cancer has been a leading cause of death in the United States with significant health care costs. Accurate prediction of cancers at an early stage and understanding the genomic mechanisms that drive cancer development are vital to the improvement of treatment outcomes and survival rates, thus resulting in significant social and economic impacts. Attempts have been made to classify cancer types with machine learning techniques during the past two decades and deep learning approaches more recently. Results In this paper, we established four models with graph convolutional neural network (GCNN) that use unstructured gene expressions as inputs to classify different tumor and non-tumor samples into their designated 33 cancer types or as normal. Four GCNN models based on a co-expression graph, co-expression+singleton graph, protein-protein interaction (PPI) graph, and PPI+singleton graph have been designed and implemented. They were trained and tested on combined 10,340 cancer samples and 731 normal tissue samples from The Cancer Genome Atlas (TCGA) dataset. The established GCNN models achieved excellent prediction accuracies (89.9-94.7%) among 34 classes (33 cancer types and a normal group). In silico gene-perturbation experiments were performed on four models based on co-expression graph, co-expression+singleton, PPI graph, and PPI+singleton graphs. The co-expression GCNN model was further interpreted to identify a total of 428 markers genes that drive the classification of 33 cancer types and normal. The concordance of differential expressions of these markers between the represented cancer type and others are confirmed. Successful classification of cancer types and a normal group regardless of normal tissues' origin suggested that the identified markers are cancer-specific rather than tissue-specific. Conclusion Novel GCNN models have been established to predict cancer types or normal tissue based on gene expression profiles. We demonstrated the results from the TCGA dataset that these models can produce accurate classification (above 94%), using cancer-specific markers genes. The models and the source codes are publicly available and can be readily adapted to the diagnosis of cancer and other diseases by the data-driven modeling research community.

The basic Generative Adversarial Networks (GAN) model is composed of the input vector, generator, and discriminator. Among them, the generator and discriminator are implicit function expressions, usually implemented by deep neural networks. GAN can learn the generative model of any data distribution through adversarial methods with excellent performance. It has been widely applied to different areas since it was proposed in 2014. In this review, we introduced the origin, specific working principle, and development history of GAN, various applications of GAN in digital image processing, Cycle-GAN, and its application in medical imaging analysis, as well as the latest applications of GAN in medical informatics and bioinformatics.

Research on document-level Neural Machine Translation (NMT) models has attracted increasing attention in recent years. Although the proposed works have proved that the inter-sentence information is helpful for improving the performance of the NMT models, what information should be regarded as context remains ambiguous. To solve this problem, we proposed a novel cache-based document-level NMT model which conducts dynamic caching guided by theme-rheme information. The experiments on NIST evaluation sets demonstrate that our proposed model achieves substantial improvements over the state-of-the-art baseline NMT models. As far as we know, we are the first to introduce theme-rheme theory into the field of machine translation.

Sign language translation (SLT) is an important application to bridge the communication gap between deaf and hearing people. In recent years, the research on the SLT based on neural translation frameworks has attracted wide attention. Despite the progress, current SLT research is still in the initial stage. In fact, current systems perform poorly in processing long sign sentences, which often involve long-distance dependencies and require large resource consumption. To tackle this problem, we propose two explainable adaptations to the traditional neural SLT models using optimized tokenization-related modules. First, we introduce a frame stream density compression (FSDC) algorithm for detecting and reducing the redundant similar frames, which effectively shortens the long sign sentences without losing information. Then, we replace the traditional encoder in a neural machine translation (NMT) module with an improved architecture, which incorporates a temporal convolution (T-Conv) unit and a dynamic hierarchical bidirectional GRU (DH-BiGRU) unit sequentially. The improved component takes the temporal tokenization information into consideration to extract deeper information with reasonable resource consumption. Our experiments on the RWTH-PHOENIX-Weather 2014T dataset show that the proposed model outperforms the state-of-the-art baseline up to about 1.5+ BLEU-4 score gains.

Currently, most work on comparing differences between simplified and traditional Chinese only focuses on the character or lexical level, without taking the global differences into consideration. In order to solve this problem, this paper proposes to use complex network analysis of word co-occurrence networks, which have been successfully applied to the language analysis research and can tackle global characters and explore the differences between simplified and traditional Chinese. Specially, we first constructed a word co-occurrence network for simplified and traditional Chinese using selected news corpora. Then, the complex network analysis methods were performed, including network statistics analysis, kernel lexicon comparison, and motif analysis, to gain a global understanding of these networks. After that, the networks were compared based on the properties obtained. Through comparison, we can obtain three interesting results: first, the co-occurrence networks of simplified Chinese and traditional Chinese are both small-world and scale-free networks. However, given the same corpus size, the co-occurrence networks of traditional Chinese tend to have more nodes, which may be due to a large number of one-to-many character/word mappings from simplified Chinese to traditional Chinese; second, since traditional Chinese retains more ancient Chinese words and uses fewer weak verbs, the traditional Chinese kernel lexicons have more entries than the simplified Chinese kernel lexicons; third, motif analysis shows that there is no difference between the simplified Chinese network and the corresponding traditional Chinese network, which means that simplified and traditional Chinese are semantically consistent.

Inter-core crosstalk causes severe signal impairment in Space Division Multiplexing Elastic Optical Networks (SDM-EON) with multi-core fibre (MCF). In order to mitigate the influence of crosstalk between cores, bidirectional MCF is used. SDM-EON has dramatically increased the network bandwidth and is the mainstream trend of future communications, but the overall cost of such an SDM-EON will increase with traditional hard-connected node structures. On-demand architecture node (AoD) in SDM-EON can dynamically select modes based on incoming services with less cost than traditional node. Although bidirectional multi-core optical fibres can mitigate the impact of crosstalk between adjacent cores, it is still necessary to develop appropriate routing, spectrum, and core assignment strategies to reduce the impact of crosstalk. In addition, the resource allocation strategy will affect the type of on-demand architecture node. This paper develops a core selection method for classified services on a multi-dimensional optical network with bidirectional multi-core fibres. It also proposes an on-demand architecture node suitable for this services classification scheme. According to the proposed core classification method, a routing and resource allocation algorithm that considers the impact of spectrum fragmentation and crosstalk between cores is proposed. The final simulation results prove that the algorithm is beneficial to improve network performance.

The dominant Neural Machine Translation (NMT) models usually resort to word-level modeling to embed input sentences into semantic space. However, it may not be optimal for the encoder modeling of NMT, especially for languages where tokenizations are usually ambiguous: On one hand, there may be tokenization errors which may negatively affect the encoder modeling of NMT. On the other hand, the optimal tokenization granularity is unclear for NMT. In this paper, we propose lattice-to-sequence attentional NMT models, which generalize the standard Recurrent Neural Network (RNN) encoders to lattice topology. Specifically, they take as input a word lattice which compactly encodes many tokenization alternatives, and learn to generate the hidden state for the current step from multiple inputs and hidden states in previous steps. Compared with the standard RNN encoder, the proposed encoders not only alleviate the negative impact of tokenization errors but are more expressive and flexible as well for encoding the meaning of input sentences. Experimental results on both Chinese–English and Japanese–English translations demonstrate the effectiveness of our models.

Bilingual word embeddings (BWEs) have proven to be useful in many cross-lingual natural language processing tasks. Previous studies often require bilingual texts or dictionaries that are scarce resources. As a result, in these studies, the exploited explicit semantic information, such as monolingual word co-occurrences and cross-lingual semantic equivalences, is often insufficient for BWE learning, leading to the limitation of learned word representations. To overcome this problem, in this paper, we study how to exploit implicit semantic constraints for better BWEs. Concretely, we first discover implicit monolingual word-level semantic equivalences by pivoting their translations in the other language. Then, we perform BWE learning under various semantic constraints. Experimental results on machine translation and cross-lingual document classification demonstrate the effectiveness of our model.

Recognizing implicit discourse relations is an important but challenging task in discourse understanding. To alleviate the shortage of labeled data, previous work automatically generates synthetic implicit data (SynData) as additional training data, by removing connectives from explicit discourse instances. Although SynData has been proven useful for implicit discourse relation recognition, it also has the meaning shift problem and the domain problem. In this paper, we first propose to use bilingually-constrained synthetic implicit data (BiSynData) to enrich the training data, which can alleviate the drawbacks of SynData. Our BiSynData is constructed from a bilingual sentence-aligned corpus according to the implicit/explicit mismatch between different languages. Then we design a multi-task neural network model to incorporate our BiSynData to benefit implicit discourse relation recognition. Experimental results on both the English PDTB and Chinese CDTB data sets show that our proposed method achieves significant improvements over baselines using SynData.

Implicit discourse relation recognition aims to discover the semantic relation between two sentences where the discourse connective is absent. Due to the lack of labeled data, previous work tries to generate additional training data automatically by removing discourse connectives from explicit discourse relation instances. However, using these artificial data indiscriminately has been proven to degrade the performance of implicit discourse relation recognition. To address this problem, we propose a co-training approach based on manual features and distributed features, which identifies useful instances from these artificial data to enlarge the labeled data. In addition, the distributed features are learned via recursive autoencoder based approaches, capable of capturing to some extent the semantics of sentences which is valuable for implicit discourse relation recognition. Experiment results on both the PDTB and CDTB data sets indicate that: (1) The learned distributed features are complementary to the manual features, and thus suitable for co-training. (2) Our proposed co-training approach can use these artificial data effectively, and significantly outperforms the baselines.

Statistical machine translation has made great progress in recent years, and Tibetan-Chinese machine translation has many needs. A phrase-based translation model is suitable for machine translation between Tibetan and Chinese, which have similar morphological changes. This paper studies the key technologies of phrase-based Tibetan-Chinese statistical machine translation, including phrase-translation models and reordering models, and proposes a phrase-based Tibetan-Chinese statistical machine translation prototype system. The method proposed in this paper has better accuracy than Moses, the current mainstream model, in the CWMT 2013 development set, and shows great performance improvement.

Human-computer interaction has been a hot topic in the development of robots. Gesture recognition is a significant part of it. Existing snake-like robot is controlled by the handle and joystick. It is worth studying that use gesture to control robots. Existing main vision-based gesture recognition methods are based on a two-stage method: hand detection and segmentation, feature extraction and template matching process. Inspired by the YOLO[1] algorithm, this paper proposed a one-stage gesture recognition and snake-like robot control method, based on deep learning, combined image segmentation and classification, realized end-to-end recognition and robot control. The snake-like robot can react to four types of human's gestures with a rapid processing time.

In biological systems, genes function in conjunction rather than in isolation. However, traditional single-cell RNA-seq (scRNA-seq) analyses heavily rely on the transcriptional similarity of individual genes, ignoring the inherent gene-gene interactions. Here, we present SCORE, a network-based method, which incorporates the validated molecular network features to infer cellular states. Using real scRNA-seq datasets, SCORE outperforms existing methods in accuracy, robustness, scalability, data integration and removal of batch effect. When applying SCORE to a newly generated human ileal scRNA-seq dataset, we identified several novel stem/progenitor clusters, including a Cripto-1+ cluster. Moreover, two distinct groups of goblet cells were identified and only one of them tended to secrete mucus. Besides, we found that the recently identified BEST4+OTOP2+ microfold cells also highly expressed CFTR, which is different from their colonic counterparts. In summary, SCORE enhances cellular state inference by simulating the dynamic changes of molecular networks, providing more biological insights beyond statistical interpretations.

Currently, there starts a research trend to leverage neural architecture for recommendation systems. Though several deep recommender models are proposed, most methods are too simple to characterize users' complex preference. In this paper, for a fine-grained analysis, users' ratings are explained from multiple perspectives, based on which, we propose our neural architectures. Specifically, our model employs several sequential stages to encode the user and item into hidden representations. In one stage, the user and item are represented from multiple perspectives and in each perspective, the representation of user and that of item put attentions to each other. Last, we metric the output representations from the final stage to approach the users' ratings. Extensive experiments demonstrate that our method achieves substantial improvements against baselines.

As a fundamental task of natural language processing, semantic role labeling (SRL) have attracted much attention of researchers in recent years. However, with increasing features being added into the studies, the performance growth trend of SRL is gradually slowing down. So new ways must be found to improve the performance of semantic analysis. Word sense information is useful for SRL task. But how to effectively make use of word sense information is a key issue. Referring to synergetics, we can regard semantic analysis process as competitive process of many semantics order parameters under coherent action and interactive collaboration of semantic role-related features and word sense-related features. Accordingly, we propose a semantic role labeling model with word sense information based on improved synergetic neural network (SNN). Our contributions are three-fold. Firstly, role-related features and word sense-related features are used to configure semantic order parameters of SNN. Secondly, network parameters are reconstructed which can reflect the relationship of driving and restraining each other between various linguistic features. Finally, we use an improved quantum particle swarm algorithm (QPSO) to realize the optimization of network parameter which has stronger search ability and faster convergence speed. By evaluating our model on the OntoNotes 2.0 corpus, the experiment results show the proposed model in this paper leads to a higher performance for SRL.

Sentiment analysis aims to identify the overall emotional polarity of a given text. It is a nontrivial task to perform sentiment analysis as sentiment information is crucial in many natural language processing applications. Previous n-gram features is derived from a bag-of-n-gram model which is insensitive to the order of the n-gram. To address this problem, we integrates distributed semantic features of word sequence, with fixed-size independent of the length of the word sequence. We also learn distributed semantic features of part-of-speech (POS) sequence as additional syntax-related clues to sentiment analysis. Our semantic features are able to capture both local contexts and global contexts automatically without involving comprehensive task-specific feature engineering. We validate the effectiveness of the method on our constructed sentiment dataset. Experiment results show that our method are able to improve the quality of sentiment analysis when comparing with several competitive baselines.

Correct and automatical semantic analysis has always been one of major goals in natural language understanding. However, due to the difficulties in deep semantic analysis, at present, the mainstream studies of semantic analysis are focused on semantic role labeling (SRL) and word sense disambiguation (WSD). Nowadays, these two issues are mostly considered as separate tasks. However, this approach ignores possible dependencies between them. In order to address the issue, an integrative semantic analysis model based on synergetic neural network (SNN) is proposed in this paper, which can easily express useful logic constraints between SRL and WSD. The semantic analysis process can be viewed as the competition process of semantic order parameters. The strongest order parameter will win by competition and desired semantic patterns will be recognized. There are three main innovations in this paper. First, an integrative semantic analysis model is proposed that jointly models word sense disambiguationand...

Word sense induction (WSI) is important to many natural language processing tasks because word sense ambiguity is pervasive in linguistic expressions. The majority of existing WSI algorithms are not applicable to capture both lexical semantics and syntactic relations without involving excessive task-specific feature engineering. Moreover, it remains a challenge to explore a sense clustering method which is capable of determining the number of word senses for the polysemous words automatically and properly. In this paper, we learn continuous semantic space representations for the ambiguous instances via recursive context composition, allowing us to capture lexical semantics and syntactic relations simultaneously. Using the learned representations of ambiguous instances, we further adapt rival penalization competitive learning to conduct instances based word sense clustering, allowing us to determine the number of word senses automatically. We validate the effectiveness of our method on the SEMEVAL-2010 WSI dataset. Experiment results show that our method is able to improve the quality of word sense clustering over several competitive baselines.

In order to adapt to the reform of transmission and distribution price, implement the company’s requirements of the policy of Delegate Power and Optimize Service, realize the precise investment of power grid and help the high quality development of power grid. It is urgent to explore a scientific and effective method of project schedule, assist the grid investment decision-making, and improve the level of investment plan management.

This paper considers the buffering costs of multi-path routing schemes and proposes a dynamic buffering cost-saving multi-path routing algorithm under differential delay constraint to save the buffering costs in elastic optical networks (EONs).

We propose a crosstalk-aware routing, spectrum and core allocation algorithm for Bi-directional Multi-core Fiber. Simulation results show that proposed algorithm can reduce blocking probability and improve resource utilization.

A flexible and load balancing algorithm (FLBSA) of joint multicast and anycast traffic is proposed to obtain multiple optimal RSA decisions using flexible STSD window in EONs. The reduction of blocking requests is experimentally demonstrated.

Representing knowledge as frames has a long history in artificial intelligence and computational linguistics. However, constructing frame banks that support frame-based processing is quite time-consuming, leading to the unavailability of usable frame banks for many languages, including Chinese. This paper proposed a method for deriving probabilistic semantic frames from HowNet, which is a well-known common-sense knowledge base and has been successfully used in many NLP applications. Unlike most previous HowNet-related work which focused on using HowNet as a lexico-semantic bank, this work viewed the HowNet dictionary as a semantic-annotated corpus. According to the proposed method, governor-role-dependent triples are firstly extracted from the concept definitions of the HowNet dictionary. Then, they are organized into frames by the governors and the probabilities are estimated based on maximum likelihood estimation (MLE). Finally, the probabilistic frames form a frame bank. Moreover, in order to overcome the data sparseness problem, a smoothing method based on HowNet’s taxonomy was put forward. To verify the constructed frame bank, we applied it in a task to recognize relationships between Chinese word pairs, which are extracted from the Chinese Message Structure Database of HowNet. The experimental results showed that, even without using context information, the system based on the constructed frame bank achieved an accuracy of 83.74%, which indicates the soundness of the constructed frame bank.

We introduce TSTD, a microblog dataset for Tibetan sentiment analysis which is publicly available. It consists of about 10K Tibetan microblogs and classified positive, negative and neutral. We present the properties of the dataset and run experiments for 3-way sentiment classification in Tibetan using feature-based and deep learning systems on the dataset.

Semantic role labeling (SRL) is a fundamental task in natural language processing to find a sentence-level semantic representation. The semantic role labeling procedure can be viewed as a process of competition between many order parameters, in which the strongest order parameter will win by competition and the desired pattern will be recognized. To realize the above-mentioned integrative SRL, we use synergetic neural network (SNN). Since the network parameters of SNN directly influence the synergetic recognition performance, it is important to optimize the parameters. In this paper, we propose an improved particle swarm optimization (PSO) algorithm based on log-linear model and use it to effectively determine the network parameters. Our contributions are two-folds: firstly, a log-linear model is introduced to PSO algorithm which can effectively make use of the advantages of a variety of different knowledge sources, and enhance the decision making ability of the model. Secondly, we propose an improved SNN model based on the improved PSO and show its effectiveness in the SRL task. The experimental results show that the proposed model has a higher performance for semantic role labeling with more powerful global exploration ability and faster convergence speed, and indicate that the proposed model has a promising future for other natural language processing tasks.

Word sense induction (WSI) aims to automatically identify different senses of an ambiguous word from its contexts. It is a nontrivial task to perform WSI in natural language processing because word sense ambiguity is pervasive in linguistic expressions. In this paper, we construct multi-granularity semantic spaces to learn the representations of ambiguous instances, in order to capture richer semantic knowledge during context modeling. In particular, we not only consider the semantic space of words, but the semantic space of word clusters and topics as well. Moreover, to circumvent the difficulty of selecting the number of word senses, we adapt a rival penalized competitive learning method to determine the number of word senses automatically via gradually repelling the redundant sense clusters. We validate the effectiveness of our method on several public WSI datasets and the results show that our method is able to improve the quality of WSI over several competitive baselines.

Measuring semantic similarity is essential to many natural language processing (NLP) tasks. One widely used method to evaluate the similarity calculating models is to test their consistency with humans using human-scored gold-standard datasets, which consist of word pairs with corresponding similarity scores judged by human subjects. However, the descriptions on how such datasets are constructed are often not sufficient previously. Many problems, e.g. how the word pairs are selected, whether or not the scores are reasonable, etc., are not clearly addressed. In this paper, we proposed a multidisciplinary method for building and validating semantic similarity standard datasets, which is composed of 3 steps. Firstly, word pairs are selected based on computational linguistic resources. Secondly, similarities for the selected word pairs are scored by human subjects. Finally, Event-Related Potentials (ERPs) experiments are conducted to test the soundness of the constructed dataset. Using the proposed method, we finally constructed a Chinese gold-standard word similarity dataset with 260 word pairs and validated its soundness via ERP experiments. Although the paper only focused on constructing Chinese standard dataset, the proposed method is applicable to other languages.

We introduce a simple and effective method to learn discourse-specific word embeddings (DSWE) for implicit discourse relation recognition. Specifically, DSWE is learned by performing connective classification on massive explicit discourse data, and capable of capturing discourse relationships between words. On the PDTB data set, using DSWE as features achieves significant improvements over baselines.