Search

This paper describes a method and technology for processing natural language texts and extracting data from the text that correspond to the semantics of an ontological model. The proposed method is distinguished by the use of a Large Language Model algorithm for text analysis. The extracted data are stored in an intermediate format, after which individuals and properties that reflect the specified semantics are programmatically created in the ontology. The proposed technology is implemented using the example of an ontological model that describes the geographical configuration and administrative–territorial division of Kazakhstan. The proposed method and technology can be applied in any subject areas for which ontological models have been developed. The results of the study can significantly improve the efficiency of using knowledge bases based on semantic networks by converting texts in natural languages into semantically linked data. [ABSTRACT FROM AUTHOR]

Semantic communication and spectrum sharing are pivotal technologies in addressing the perennial challenge of scarce spectrum resources for the sixth-generation (6G) communication networks. Notably, scant attention has been devoted to investigating semantic resource allocation within spectrum sharing semantic communication networks, thereby constraining the full exploitation of spectrum efficiency. To mitigate interference issues between primary users and secondary users while augmenting legitimate signal strength, the introduction of Intelligent Reflective Surfaces (IRS) emerges as a salient solution. In this study, we delve into the intricacies of resource allocation for IRS-enhanced semantic spectrum sharing networks. Our focal point is the maximization of semantic spectral efficiency (S-SE) for the secondary semantic network while upholding the minimum quality of service standards for the primary semantic network. This entails the joint optimization of parameters such as semantic symbol allocation, subchannel allocation, reflective coefficients of IRS elements, and beamforming adjustment of secondary base station. Recognizing computational intricacies and interdependence of variables in the non-convex optimization problem formulated, we present a judicious approach: a hybrid intelligent resource allocation approach leveraging dueling double-deep Q networks coupled with the twin-delayed deep deterministic policy. Simulation results unequivocally affirm the efficacy of our proposed resource allocation approach, showcasing its superior performance relative to baseline schemes. Our approach markedly enhances the S-SE of the secondary network, thereby establishing its prowess in advancing the frontiers of semantic spectrum sharing (S-SE). [ABSTRACT FROM AUTHOR]

This article aims to evaluate how teachers, in training and in-service, define the concepts of randomness, probability, chance and variability, fundamental terms in the teaching of statistics. To this end, a printed recording protocol of natural semantic networks was applied to 16 teachers in training and 58 in-service teachers, selected through non-probabilistic sampling. The results provide evidence that the concept with the lowest conceptual density in both groups is variability. Likewise, a greater presence of similar words was observed between randomness, chance and probability, despite being different concepts. Another finding is the association of the inducing concepts to words that are used within the basic lexicon in Chilean Spanish, whose dictionary is not specialized in statistics. It is concluded about the scarce use of technical language by the participants, which would eventually affect the teaching of statistics. [ABSTRACT FROM AUTHOR]

The impact of COVID-19 pandemic is farreaching, encompassing the social, economic, and psychological aspects of the society. . In order to prevent the spread of COVID-19, various countries, including South Korea, have entered into long-term home care and distance learning systems. However, distance learning experiments conducted in many countries have raised whether face-to-face education can be replaced by distance learning. Therefore, this study analyzed public opinion, social perception, and field trends based on media reports on remote learning. For this purpose, 2,600 articles from 11 newspapers and four broadcasters related to remote learning were collected in this study. Based on this data, keyword trend analysis, topic modeling analysis, semantic network analysis, and emotion analysis were performed. This study provides valuable insights into the impacts and perceptions of remote learning during the COVID-19 pandemic. It highlights the complexities of remote education, from influencing educational policies to affecting students' and parents' emotional states. The findings underscore the potential of remote learning in driving educational innovation while also pointing out its role in widening educational disparities. This research offers essential perspectives for shaping future educational strategies in crisis situations, emphasizing a balanced approach to harness the benefits of remote learning while mitigating its challenges. [ABSTRACT FROM AUTHOR]

This research focuses on developing effective algorithms for semantic knowledge graph searches in the context of finding causal answers, which is highly relevant due to the widespread use of semantic networks. The study's primary goal is to examine the construction principles of a semantic knowledge graph search model tailored for causal answers, with a focus on deep neural semantic search characteristics. The research methodology combines system analysis methods for building knowledge graphs with an analytical investigation of deep neural semantic search aspects. The results provide insights into the construction of a semantic knowledge graph search model for causal answers, addressing the challenges and methodologies involved in building such a model. It underscores the significance of knowledge graphs in modern information systems and their potential applications in various domains. However, further scientific research is essential to explore the practical applications of deep neural semantic search knowledge graphs in various information systems, which are used across different aspects of everyday life. The practical significance of this research extends to various applications in information retrieval, knowledge management, and problem-solving, making it a valuable contribution to the advancement of technology and understanding of natural language text. [ABSTRACT FROM AUTHOR]

Traditional methods of weakly supervised semantic segmentation (WSsegmentation) for point cloud scenes have several limitations including limited precision and difficulty in handling complex scenes due to imprecise labels or partial annotations. To address these issues, we perform view-based adversarial training on the original point cloud scene samples through view resampling and Gaussian noise perturbation to reduce overfitting. Combining a self-attention mechanism with multi-layer perceptrons and point cloud segmentation strategy, we can perform dimensionality enhancement and dimensionality reduction operations to better capture the local features of point cloud data. Finally, we can obtain the semantic segmentation results of point cloud scene by fusing local and global semantic features. In the design of the network loss function, we combine the Siamese loss and the smoothness loss and the cross-entropy loss to improve the ability and fidelity of semantic networks. Specifically, the Siamese loss is used to compute the distance between different augmented point cloud data in their feature embedding space and the smoothness loss is used to penalize the discontinuity of semantic information between adjacent regions. The proposed weakly supervised segmentation network achieves an overall segmentation accuracy close to fully supervised segmentation methods and outperforms most of the existing weakly supervised segmentation methods by 5% to 10% in scene segmentation in terms of mIoU on S3DIS, ShapeNet, and PartNet datasets. Extensive experiments demonstrate the robustness, effectiveness, and generalization of the proposed point cloud segmentation network. [Display omitted] • We propose effective data augmentation methods, including view resampling and adversarial training, to tackle partial annotations. • To fully integrate the global and local features of point cloud, we employ high-dimensional feature self-attention mechanism and multi-layer perceptrons to fuse multi-level features. • We present a novel point cloud segmentation framework, which can achieve the higher segmentation accuracy than traditional segmentation networks for different point cloud scenes. • Extensive experiments demonstrate the robustness, effectiveness, and generalization of the proposed weakly supervised point cloud semantic segmentation network. [ABSTRACT FROM AUTHOR]

As a learning method of heterogeneous graph representation, heterogeneous graph neural networks can effectively extract complex structural and semantic information from heterogeneous graphs, and perform excellently in node classification and link prediction tasks to provide strong support for the representation and analysis of knowledge graphs. Due to the existence of some noisy interactions or missing interactions in the heterogeneous graphs, the heterogeneous graph neural network incorporates erroneous neighbor features, thus affecting the overall performance of the model. To solve the above problems, in this paper we proposes a heterogeneous graph structure learning model enhanced by multi-view contrast. Firstly, the semantic information in the heterogeneous graph is maintained by the meta-path, and the similarity graph is generated by calculating the feature similarity among the nodes under each meta-path, which is fused with the meta-path graph to optimize the graph structure. By contrasting the similarity graph and meta-path graph as multiple views, the graph structure is optimized without supervision information, and the dependence on supervision signals is eliminated. Finally, for addressing the problem that the learning ability of the neural network model is insufficient at the initial training stage and there are often erroneous interactions in the generated graph structure, we design a progressive graph structure fusion method. Through incrementalweighted addition of meta-path graphs and similarity graphs, theweight of similarity graphs in the fusion is changed. This not only prevents erroneous interactions from being introduced in the initial training stage but also achieves the purpose of employing the interactions in similarity graphs to suppress interference interactions or complete missing interactions, which leads to the optimized heterogeneous structure. Meanwhile, node classification and node clustering are selected as the verification tasks of graph structure learning. The experimental results on four real heterogeneous graph datasets prove that the proposed learning method is feasible and effective. Compared with the optimal comparison model, the performance of this model has been significantly improved under both tasks. [ABSTRACT FROM AUTHOR]

This paper uses a new textual data index for predicting stock market data. The index is applied to a large set of news to evaluate the importance of one or more general economic‐related keywords appearing in the text. The index assesses the importance of the economic‐related keywords, based on their frequency of use and semantic network position. We apply it to the Italian press and construct indices to predict Italian stock and bond market returns and volatilities in a recent sample period, including the COVID‐19 crisis. The evidence shows that the index captures the different phases of financial time series well. Moreover, results indicate strong evidence of predictability for bond market data, both returns and volatilities, short and long maturities, and stock market volatility. [ABSTRACT FROM AUTHOR]

This research studies the impact of online news on social and economic consumer perceptions through semantic network analysis. Using over 1.8 million online articles on Italian media covering four years, we calculate the semantic importance of specific economic-related keywords to see if words appearing in the articles could anticipate consumers' judgments about the economic situation and the Consumer Confidence Index. We use an innovative approach to analyze big textual data, combining methods and tools of text mining and social network analysis. Results show a strong predictive power for the judgments about the current households and national situation. Our indicator offers a complementary approach to estimating consumer confidence, lessening the limitations of traditional survey-based methods. [ABSTRACT FROM AUTHOR]

It is crucial to observe how international policy initiatives are implemented in each country's environmental policies to ensure global environmental conservation and sustainable development. This study aims to comprehensively analyze the trend of environmental policy in Korea over time. To this aim, we use semantic network analysis to explore the main keywords and to identify the characteristics of the key themes in each period based on the national environmental comprehensive plan. Our results show that the value of economic growth and technological development are regarded as priorities in Korean environmental policy because the economy and technology are closely related to the environment. Similarly, when it comes to the fundamental concept of environmental policy, the concept of "green growth" is applied in preference to the concept of "sustainability." The results contribute to expanding the approach of existing policy research by providing the potential for semantic network analysis based on policy documents. [ABSTRACT FROM AUTHOR]

Social media platforms such as Twitter provide opportunities for governments to connect to foreign publics and influence global public opinion. In the current study, we used social and semantic network analysis to investigate China's digital public diplomacy campaign during COVID-19. Our results show that Chinese state-affiliated media and diplomatic accounts created hashtag frames and targeted stakeholders to challenge the United States or to cooperate with other countries and international organizations, especially the World Health Organization. Telling China's stories was the central theme of the digital campaign. From the perspective of social media platform affordance, we addressed the lack of attention paid to hashtag framing and stakeholder targeting in the public diplomacy literature. [ABSTRACT FROM AUTHOR]

To study the international news reporting patterns of the Chinese media, this research provides a longitudinal analysis of the news coverage of Sino-Japan summit meetings from 1972 to 2019 in the People's Daily. Based on the framing theory, the results of the analysis show that different aspects of the summit are selected to be covered in the news depending on the Chinese government's policy towards Japan in different time periods. However, the news coverage itself has maintained a high degree of consistency. Before the period of reform and the opening up of China, there was no mention of the summit meetings' content; reports only introduced the date, location and attendees. After the reform, information reported about the summit meetings increased and the coverage began including quotations from both the Chinese and Japanese heads of state; at the same time, reports adopted a 'propagandized' text editing mode by quoting more from Chinese leaders than Japanese leaders. Furthermore, the People's Daily framed the Chinese government's attempts to stabilize and develop Sino-Japanese relations over time. [ABSTRACT FROM AUTHOR]

This article examines the semantic development of significant keywords in Taiwanese cultural discourse from a corpus-based inductive perspective. It contextualises these keywords ideologically and socially via extended network analysis and innovatively also via sociological analysis. The research investigated lexical patterns in what we refer to as the Taiwan Early Post-war Corpus, which consists of culturally oriented articles from 1945 to 1949. It treated shared ideologically loaded keywords as indicators of different ideologies which in turn identify the social groups who propagate them. Through this process we discovered three ideologically inclined semantic fields, which emphasise connections to the dominant cultural discourses of the central and local governments, as well as one that emphasises Taiwanese subjectivity. To understand their contemporaneous dynamics, the study examined networks associated with the authors and periodicals through shared keywords. Using the Taiwan Biographical Ontology, the study complemented ideological patterns with positional analyses of authors' social involvements, which allowed us to treat them as proxies for various types of social, cultural, economic and political capital. We could thus characterise the habitus of each cluster and its position on an ideological map, thereby creating multi-layered networks of concepts, ideologies and social patterns. [ABSTRACT FROM AUTHOR]

Classifications are widely used in semantic networks and decision support systems based on formal knowledge and are part of computer ontologies. Classifications and computer ontologies built on them are the result of the work of one or more experts. As a result, such classifications reflect the subjective view of the author or authors on the world and the relationship between the classes (concepts) of the created classification. In the work, the authors propose an approach that will allow assessing how correctly the classification is constructed. [ABSTRACT FROM AUTHOR]

By capturing the online evaluation data of tourists on Zhuhai coastal tourist attractions in the two domestic OTA platforms Ctrip and Mafengwo, word cloud and word frequency statistics and complex semantic network analysis are used to carry out the research of image perception and emotional evaluation. The results show that: (1) Zhuhai City is the main body of the composition of tourists' image perception of Zhuhai coastal tourism, showing the 3S (seafood, sand, scenery) coastal tourism image of Zhuhai characteristics, and the rapid development of Zhuhai City is an important driving force for the transformation of the city's image perception. (2) The single core and two wings is the hierarchical pattern of tourists' image perception of Zhuhai's coastal tourism, namely Zhuhai as the centre and sand and seafood as the two wings. (3) After evaluating the emotions of tourists in Zhuhai's coastal area, it is found that overall the tourists' perceived image is relatively good, with a large proportion of positive emotions, but there are also some negative and neutral emotions. [ABSTRACT FROM AUTHOR]

This contribution addresses the general theme of Multifunctionality and syncretism in non-finite forms. When we see a particular form in a number of different environments yielding different effects, what does it mean to say that this reflects multifunctionality rather than syncretism? As a refinement of multifunctionality the notion of 'multi-effectuality' is introduced: one element participates in different derivations, performing one and the same operation, but yields different interpretive effects depending on the environment. The analysis takes as a starting point the role of the affix -en in Dutch. The canonical view is that when attached to a verb stem, -en functions either as an infinitival marker or as a nominalizing affix. It is shown that this view cannot be maintained. The "functions" of -en are unified by analyzing it as projecting an 'eventuality expression', with different effects depending on the syntactic environment. The analysis is extended to English, focusing on the contrasts between Dutch and English infinitives and on the role of the English -ing affix in gerunds and participles. The discussion proceeds with an analysis of Frisian with its two forms of the infinitive and concludes with some reflections on the relation between syntactic and semantic categories. [ABSTRACT FROM AUTHOR]

Semantic network is a network of concepts connected by semantic relations. It contains two forms of binary semantic network and multiplex semantic network. The associated fact prediction is a link prediction task that aims to infer the implicitly connected facts by mining the high-level representation of the network. Previous methods for associated fact prediction put much emphasis on the topological feature of network but not utilize the information of semantic expression. This paper proposes a Semantic Network Encoder (SemNE), which learns a feature mapping function from the binary semantic networks and can be applied to the multiplex semantic networks in a pre-training manner. SemNE is a two-stage framework that contains an embedding encoder and a prediction decoder. It jointly models the semantic information and network topology to enrich the network representation. A word self-organization method based on the factual boundary is proposed to unify the topological feature and the semantic feature representations. Experimental results on binary semantic networks show that SemNE achieves the state-of-the-art results in associated fact prediction and experimental results on multiplex semantic networks show that SemNE is scalable and can effectively improve the performance of existing models. [ABSTRACT FROM AUTHOR]

Recent years have seen a significant and rapid rise in the development and deployment of many urban digital twins worldwide. Urban digital twins provide a central platform for incorporating data and knowledge from different sources and fields, and can thus be used for many urban-related processes such as urban planning, analysis, monitoring and visualization. One of the key requirements of digital twins in general and urban digital twins in particular is the continuous, bidirectional data flow between the physical entity and its digital counterpart. Consequently, this requires the ability to both detect and understand changes between different temporal versions of the physical and digital entities. In the context of smart cities and semantic 3D city models however, only a few studies have addressed this so far. This is due to the facts that: (1) Semantic 3D city models (mostly encoded in CityGML) contain multifaceted information modelled in a complex inheritance hierarchy, which complicates their change detection process; (2) As cities constantly evolve over time, matching their datasets often results in a large number of changes, which must be further analysed and processed to produce meaningful information; (3) Individual changes in the datasets are multi-layered as well as often correlated and cannot be fully understood without considering their context in the city model; and (4) Different types of changes are perceived and interpreted differently by different stakeholders, meaning that the outcome of the change detection and interpretation process must ultimately serve the human factor. Therefore, to address these challenges, this research proposes a Path-tracing Semantic Network (PSN) to interpret detected changes in semantic 3D city models. The framework represents the multidimensional nature of changes together with stakeholders in a semantic network, where their interrelations can be analysed explicitly using graph-based path-tracing methods. [ABSTRACT FROM AUTHOR]

China's state-run media are the mouthpiece of the government. During public health crises such as the COVID-19 pandemic, they are responsible for disseminating essential information to the public on behalf of the government. This study examined the Sina Weibo posts published by three leading state-run media entities (CCTV, People's Daily, and Xinhua News Agency) during the first wave of the COVID-19 outbreak. Semantic networks were extracted from posts during each stage of the outbreak, and clusters of nodes representing communication themes were identified, including investigations of the coronavirus, governmental policies and response efforts, case updates, prevention and control, and medical treatment. These themes indicate the use of information and bolstering strategies to maintain and increase government legitimacy. The findings can inform future risk and crisis communication during public health crises. [ABSTRACT FROM AUTHOR]

Despite their drawback, multiple-choice questions are an enduring feature in instruction because they can be answered more rapidly than open response questions and they are easily scored. However, it can be difficult to generate good incorrect choices (called "distractors"). We designed an algorithm to generate distractors from a semantic network for four types of multiple choice questions in biology. By recruiting 200 participants from Amazon Mechanical Turk, the machine-generated distractors were compared to human-generated distractors in terms of question difficulty, question discrimination and distractor usefulness. The machine-generated and human-generated distractors performed very closely on all the three measures, suggesting that generating distractors from a semantic network for simple multiple choice questions is a viable method. [ABSTRACT FROM AUTHOR]

To accurately and formally represent the historical trajectory and present the current situation of land use/land cover (LULC), numerous types of classification standards for LULC have been developed by different nations, institutes, organizations, etc.; however, these land cover classification systems and legends generate polysemy and ambiguity in integration and sharing. The approaches for dealing with semantic heterogeneity have been developed in terms of semantic similarity. Generally speaking, these approaches lack domain ontologies, which might be a significant barrier to implementing these approaches in terms of semantic similarity assessment. In this paper, we propose an ontological approach to assess the similarity of the domain of LULC classification systems and standards. We develop domain ontologies to explicitly define the descriptions and codes of different LULC classification systems and standards as semantic information, and formally organize this semantic information as rules for logical reasoning. Then, we utilize a Bayes algorithm to create a conditional probabilistic model for computing the semantic similarity of terms in two separate LULC land cover classification systems. The experiment shows that semantic similarity can be effectively measured by integrating a probabilistic model based on the content of ontology. [ABSTRACT FROM AUTHOR]

Knowledge graphs are large-scale semantic networks that considerably impact knowledge representation. Mining hidden knowledge from existing data, including triplet knowledge reasoning, is a primary objective of knowledge graphs. With the development of Neural Network (NN) and Deep Learning (DL), the interpretability of triplet knowledge reasoning gradually decreases; furthermore, what machines learn is not actual reasoning but digital reasoning shortcuts. To solve this problem, more background knowledge needs to be introduced into knowledge graphs: causal graphs can offer valuable causal logic knowledge for reasoning; temporal quadruples can provide essential temporal distribution details; and commonsense graphs can furnish pertinent commonsense understanding to support reasoning. In recent years, many scholars have incorporated additional background knowledge into knowledge graphs to construct more complex reasoning mechanisms. This paper reviews the basic concepts and definitions of knowledge reasoning and the reasoning methods used for knowledge graphs. Specifically, we dissect the reasoning methods into four categories: triplet reasoning, causal inference, temporal inference, and commonsense reasoning. Finally, we discuss the remaining challenges and research opportunities related to knowledge graph reasoning. [ABSTRACT FROM AUTHOR]

To address semantic misperception caused by distortion and accurately simulate human perceptual processes, this study proposed a method that utilised dual-stream mutually adaptive feature mapping. To extract complementary features, a dual-stream unsupervised method was used during the training phase. One stream was responsible for the extraction of low-level and global features, whereas the other was dedicated to extracting high-level semantic and positional features. Following the freezing of feature extraction network weights, we proposed a structure that used standard deviation labels to predict the quality distribution. The experimental results obtained from 10 published image quality databases demonstrated the superiority of the proposed algorithm. The algorithm outperformed the majority of mainstream methods when evaluated on authentic distortion databases and exhibited competitive performance on synthetic distortion databases. [ABSTRACT FROM AUTHOR]

Traditional supervised techniques rely on labeled data, which are not available for unknown classes. Generalized Category Discovery (GCD) aims to categorize data into both known and unknown classes. We proposed a method, Unknown Sample Selection and Discriminative Classifier Learning for GCD (USSDCL). Firstly, we map the features to a linear subspace to identify dimensions that best represent the semantic differences among the known classes and introduce a neighborhood relevance score to assess the consistency of labels among neighbors. Using this score, we can efficiently distinguish between known, unknown, and uncertain samples. Next, we leverage both the known samples, pseudo-labeled unknown samples, and the uncertain samples to train a discriminative classifier. This classifier incorporates both a cross-entropy loss and an uncertainty augmentation loss to encourage closer predictions between uncertain samples and their augmentations, enhancing the classifier's discriminative capability. Our model exhibits enhanced performance in classifying both known and unknown samples. [ABSTRACT FROM AUTHOR]

According to the online database landmarkmap, up to an estimated 50% or more of the world’s habitable land is held by indigenous peoples and communities. While legal and procedural provisions are being made for bureaucratically managing the many different types of tenure relations in this domain, there continues to be a lack of tools and expertise needed to quickly and accurately document customary and indigenous land rights. Software and hardware tools that have been designed for documenting land tenure through communities continue to assume a parcel-based model of land as well as categories of land relations (RRR) largely dimensionally similar to statutory land rights categories. The SmartSkeMa approach to land tenure documentation combines sketching by hand with aerial imagery and an ontology-based model of local rules regulating land tenure relations to produce a system specifically designed to allow accurate documentation of land tenure from a local perspective. In addition, the SmartSkeMa adaptor which is an OWL-DL based set of rules for translating local land related concepts to the LADM concepts provides a more high-level view of the data collected (i.e., what does this concept relate to within the national LADM profile?) In this paper we present the core functionalities of SmartSkeMa using examples from Kenya and Ethiopia. Based on an expert survey and focus groups held in Kenya, we also analyze how the approach fairs on the Fit-for-Purpose Land Administration tools scale. The results indicate that the approach could be beneficial in scaling up mapping of community and customary lands as well as help reduce conflict through its participatory nature. [ABSTRACT FROM AUTHOR]

The dominant video captioning methods employ the attentional encoder–decoder architecture, where the decoder is an autoregressive structure that generates sentences from left‐to‐right. However, these methods generally suffer from the exposure bias issue and neglect the guidance of future output contexts obtained from the right‐to‐left decoding. Here, the authors propose a new symmetric bidirectional decoder for video captioning. The authors first integrate the self‐attentive multi‐head attention and bidirectional gated recurrent unit for capturing the long‐term semantic dependencies in videos. The authors then apply one single decoder to generate accurate descriptions from left‐to‐right and right‐to‐left simultaneously. The decoder in each decoding direction performs two cross‐attentive multi‐head attention modules to consider both the past hidden states from the same decoding direction and the future hidden states from the reverse decoding direction at each time step. A symmetric semantic‐guided gated attention module is specially devised to adaptively suppress the irrelevant or misleading contents in the past or future output contexts and retain the useful ones for avoiding under‐description. Experimental evaluations on two widely applied benchmark datasets: Microsoft research video to text and Microsoft video description corpus, demonstrate that the authors' proposed method obtains substantially state‐of‐the‐art performance, which validates the superiority of the bidirectional decoder. [ABSTRACT FROM AUTHOR]

Source code suggestion is the utmost helpful feature in the integrated development environments that helps to quicken software development by suggesting the next possible source code tokens. The source code contains useful semantic information but is ignored or not utilised to its full potential by existing approaches. To improve the performance of source code suggestion, the authors propose a deep semantic net (DeepSN) that makes use of semantic information of the source code. First, DeepSN uses an enhanced hierarchical convolutional neural network combined with code‐embedding to automatically extract the top‐notch features of the source code and to learn useful semantic information. Next, the source code's long and short‐term context dependencies are captured by using long short‐term memory. We extensively evaluated the proposed approach with three baselines on ten real‐world projects and the results are suggesting that the proposed approach surpasses state‐of‐the‐art approaches. On average, DeepSN achieves 7.6% higher accuracy than the best baseline. [ABSTRACT FROM AUTHOR]

The semantic network structure is a core aspect of the mental lexicon and is, therefore, a key to understanding language development processes. This study investigated the structure of the semantic network of adolescents with intellectual disability (ID) and children with typical development (TD) using network analysis. The semantic networks of the participants (nID = 66; nTD = 49) were estimated from the semantic verbal fluency task with the pathfinder method. The groups were matched on the number of produced words. The average shortest path length (ASPL), the clustering coefficient (CC), and the network's modularity (Q) of the two groups were compared. A significantly smaller ASPL and Q and a significantly higher CC were found for the adolescents with ID in comparison with the children with TD. Reasons for this might be differences in the language environment and differences in cognitive skills. The quality and quantity of the language input might differ for adolescents with ID due to differences in school curricula and because persons with ID tend to engage in different out-of-school activities compared to TD peers. Future studies should investigate the influence of different language environments on the language development of persons with ID. [ABSTRACT FROM AUTHOR]

The “Growing Up Asian American” (GUAA) tag on YouTube conveys the marginalized experiences of Asian American women and offers narratives not shown in the mainstream media as a networked counterpublic. By exploring YouTube videos using this tag, I analyzed counterpublic discourses formed by nonstar Asian American YouTubers and how these discourses were shaped by the culture and medium of YouTube. Using mixed methods, including semantic network analysis and discursive textual analysis, I explored the unique traits of the GUAA Asian American networked counterpublic on YouTube. Although the discourses formed around the tag are considered relatively soft, an issue compounded by the nature of microcelebrity culture, the GUAA tag has also allowed Asian Americans to narrate their own personal and affective stories. The audiovisual nature of YouTube allows Asian Americans to talk with their bodies and voices to express their rich, heterogeneous, and intersectional identity, which cannot be reduced to one monolithic media portrayal. [ABSTRACT FROM AUTHOR]

The months-long anti-extradition bill movement in Hong Kong has gained worldwide attention. Grounded in the network agenda-setting (NAS) model, this study utilizes a machine-learning approach to analyze the related coverage of mainstream newspapers in Hong Kong, Mainland China, the U.S. and the U.K. (N = 2118), as well as discussions on Twitter (N = 152,509). Network visualizations showed that each media utilized a unique approach to highlight and connect the substantive and affective attributes. Time-series analyses revealed an overall reciprocal whilst asymmetrical association between the newspapers and Twitter, in which the latter exhibited a stronger influence on the former, particularly in terms of substantive attribute agendas. Yet, Twitter's impact shrank in terms of the affective attribute agendas and the NAS models. Newspapers, though exerted rather limited impact on Twitter, maintained a certain extent of independence in setting their affective attribute agendas and NAS models. This study enriches the NAS literature through combining substantive and affective attributes in semantic networks. [ABSTRACT FROM AUTHOR]

With various structured data, such as the company size, loan balance, and savings accounts, the voice of customer (VOC), which is text data containing contact history and counseling details was analyzed in this study. To analyze unstructured data, the term frequency--inverse document frequency (TF-IDF) analysis, semantic network analysis, sentiment analysis, and a convolutional neural network (CNN) were implemented. A performance comparison of the models revealed that the predictive model using the CNN provided the best performance with regard to predictive power, followed by the model using the TF-IDF, and then the model using semantic network analysis. In particular, a character-level CNN and a wordlevel CNN were developed separately, and the character-level CNN exhibited better performance, according to an analysis for the Korean language. Moreover, a systematic selection model for optimal text mining techniques was proposed, suggesting which analytical technique is appropriate for analyzing text data depending on the context. This study also provides evidence that the results of previous studies, indicating that individual customers leave when their loyalty and switching cost are low, are also applicable to corporate customers and suggests that VOC data indicating customers' needs are very effective for predicting their behavior. [ABSTRACT FROM AUTHOR]

To better solve scale variance problem, deep multi-scale methods usually detect objects of different scales by different in-network layers. However, the semantic levels of features from different layers are usually inconsistent. In this paper, we propose a multi-branch and high-level semantic network by gradually splitting a base network into multiple different branches. As a result, the different branches have same depth and the output features of different branches have similarly high-level semantics. Due to the difference of receptive fields, the different branches are suitable to detect objects of different scales. Meanwhile, the multi-branch network does not introduce additional parameters by sharing the convolutional weights of different branches. To further improve detection performance, skip-layer connections are used to add context to the branch of relatively small receptive field, and dilated convolution is incorporated to enlarge the resolutions of output feature maps. When they are embedded into Faster RCNN architecture, the weighted scores of proposal generation network and proposal classification network are further proposed. Experiments on three pedestrian datasets (i.e., the KITTI dataset, the Caltech dataset, and the Citypersons dataset), one face dataset (i.e., the WIDER FACE dataset), and two general object datasets (i.e., the COCO benchmark and the PASCAL VOC dataset) demonstrate the effectiveness and generality of proposed method. On these datasets, our method achieves state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Despite their mastery over commercially popular textbooks, many EFL learners feel communicatively incompetent in real communicative interactions outside the classroom context. This data-driven study aims at exploring the roots of this educational ill by interviewing nine experienced English language teachers teaching in private language schools of Mashhad, the capital city of Khorasan province, Iran. An initial round of data collection and analysis yielded transfer value as the missing link in ELT materials, which accounts for learners' communicative incompetence. Iterative data collection and analysis and the constant comparative technique presented a thick description of what transfer value is, related it to its underlying causes and predicted its implications for teaching, learning and materials development. Furthermore, importing the transcription of the participants' perspectives and the emerged codes and categories into MAXQDA yielded: (1) a semantic network which shows that the final conceptualisation of teachers' perspectives enjoys descriptive, explanatory and predictive adequacy; (2) a code relation browser that clearly shows the degree of overlap or concurrence of the codes; and (3) a set of charts that illustrate the degree to which the emerged categories are grounded in the participants' perspectives. The findings have clear implications for materials developers and practitioners. [ABSTRACT FROM AUTHOR]

Knowledge graph is a kind of semantic network for information retrieval. How to construct a knowledge graph that can serve the power system based on the behavior data of dispatchers is a hot research topic in the area of electric power artificial intelligence. In this paper, we propose a method to construct the dispatch knowledge graph for the power grid. By leveraging on dispatch data from the power domain, this method first extracts entities and then identifies dispatching behavior relationship patterns. More specifically, the method includes three steps. First, we construct a corpus of power dispatching behaviors by semi-automated labeling. And then, we propose a model, called the BiLSTM-CRF model, to extract entities and identify the dispatching behavior relationship patterns. Finally, we construct a knowledge graph of power dispatching data. The knowledge graph provides an underlying knowledge model for automated power dispatching and related services and helps dispatchers perform better power dispatch knowledge retrieval and other operations during the dispatch process. [ABSTRACT FROM AUTHOR]

Zero-shot sketch-based image retrieval (ZS-SBIR) is a specific cross-modal retrieval task that involves searching natural images through the use of free-hand sketches under the zero-shot scenario. Most previous methods project the sketch and image features into a low-dimensional common space for efficient retrieval, and meantime align the projected features to their semantic features (e.g., category-level word vectors) in order to transfer knowledge from seen to unseen classes. However, the projection and alignment are always coupled; as a result, there is a lack of explicit alignment that consequently leads to unsatisfactory zero-shot retrieval performance. To address this issue, we propose a novel progressive cross-modal semantic network. More specifically, it first explicitly aligns the sketch and image features to semantic features, then projects the aligned features to a common space for subsequent retrieval. We further employ cross-reconstruction loss to encourage the aligned features to capture complete knowledge about the two modalities, along with multi-modal Euclidean loss that guarantees similarity between the retrieval features from a sketch-image pair. Extensive experiments conducted on two popular large-scale datasets demonstrate that our proposed approach outperforms state-of-the-art competitors to a remarkable extent: by more than 3% on the Sketchy dataset and about 6% on the TU-Berlin dataset in terms of retrieval accuracy. [ABSTRACT FROM AUTHOR]

Are bilinguals more creative than monolinguals? Some prior research suggests bilinguals are more creative because the knowledge representations for their second language are similarly structured to those of highly creative people. However, there is contrasting research showing that the knowledge representations of bilinguals' second language are actually structured like those of less creative people. Finally, there is growing skepticism about there being differences between bilinguals and monolinguals on non-language tasks (e.g., the bilingual advantage for executive control). We tested whether bilinguals tested in their second language are more or less creative than both monolinguals and bilinguals tested in their first language. Participants also took a repeated semantic fluency test that we used to estimate individual semantic networks for each participant. We analyzed our results with Bayesian statistics and found support for the null hypothesis that bilingualism offers no advantage for creativity. Further, using best practices for estimating semantic networks, we found support for the hypothesis that there is no association between an individual's semantic network and their creativity. This is in contrast with published research, and suggests that some of those findings may have been the result of idiosyncrasies, outdated methods for estimating semantic networks, or statistical noise. Our results call into question reported relations between bilingualism and creativity, as well as semantic network structure as an explanatory mechanism for individual differences in creativity. [ABSTRACT FROM AUTHOR]

Ontologies represent the standard way to model the knowledge about specific domains. This holds also for the legal domain where several ontologies have been put forward to model specific kinds of legal knowledge. Both for standard users and for law scholars, it is often difficult to have an overall view on the existing alternatives, their main features and their interlinking with the other ontologies. To answer this need, in this paper, we address an analysis of the state-of-the-art in legal ontologies and we characterise them along with some distinctive features. This paper aims to guide generic users and law experts in selecting the legal ontology that better fits their needs and in understanding its specificity so that proper extensions to the selected model could be investigated. [ABSTRACT FROM AUTHOR]

Semantic segmentation is an important but challenging task in the field of medical image analysis. Automatic labeling for different anatomical structures can be useful for disease diagnosis, treatment planning and development/degeneration evaluation. However, due to the large shape and appearance variance among different subjects, accurate and reliable semantic segmentation is difficult for automatic labeling and time-consuming for manual labeling. In this paper, we propose a multi-class semantic segmentation algorithm based on the 3D fully convolutional networks, exploring the multimodal image blocks. The use of multimodal image blocks naturally allows an effective data augmentation. Besides, we also investigated the use of ensemble learning and conditional random fields (CRFs) as the post-processing step, which obtains the spatially consistent segmentation results. Our proposed method was evaluated on two public benchmarks: BRATS2017 and MNI-HiSUB25. For both datasets, the experimental results demonstrated that our method was able to establish an effective semantic segmentation framework with multi-modality data and achieve better performance compared to the conventional fully convolutional network. [ABSTRACT FROM AUTHOR]

Organized semantic networks reflecting distinctions within and across domains of knowledge are critical for higher-level cognition. Thus, understanding how semantic structure changes with experience is a fundamental question in developmental science. This study probed changes in semantic structure in 4-6 year-old children (N = 29) as a result of participating in an enrichment program at a local botanical garden. This study presents the first direct evidence that (a) the accumulation of experience with items in a domain promoted increases in both within- and across-domain semantic differentiation, and that (b) this experience-driven semantic differentiation generalized to nonexperienced items. These findings have implications for understanding the role of experience in building semantic networks, and for conceptualizing the contribution of enrichment experiences to academic success. [ABSTRACT FROM AUTHOR]

Crowd counting is a challenging task dealing with the variation of an object scale and a crowd density. Existing works have emphasized on skip connections by integrating shallower layers with deeper layers, where each layer extracts features in a different object scale and crowd density. However, only high-level features are emphasized while ignoring low-level features. This paper proposes an estimation network by passing high-level features to shallow layers and emphasizing its low-level feature. Since an estimation network is a hierarchical network, a high-level feature is also emphasized by an improved low-level feature. Our estimation network consists of two identical networks for extracting a high-level feature and estimating the final result. To preserve semantic information, dilated convolution is employed without resizing the feature map. Our method was tested in three datasets for counting humans and vehicles in a crowd image. The counting performance is evaluated by mean absolute error and root mean squared error indicating the accuracy and robustness of an estimation network, respectively. The experimental result shows that our network outperforms other related works in a high crowd density and is effective for reducing over-counting error in the overall case. [ABSTRACT FROM AUTHOR]

How do the semantic logics that different words accommodate in different languages map onto studies of social realities internationally and interdisciplinarily? This article is an ethnographic study of obshcheye – a corpus of phenomena pertaining to communal life in Russia. Similar to the English term 'commons', it marks the zone of the public – that which is shared and collective. In contrast to the commons, it displays greater semantic polyphony, bringing together social, discursive and affective qualities. Our analysis demonstrates that various semantic subsets of obshcheye sensitize research differently from the commons by indexing different societal concerns. They tune us into a wide set of concerns – with time (not wanting to be 'Soviet'), ownership (worrying about what is 'no one's'), affective connectivity (one sits and waits for a conversation), and the act of caring for people and for spaces. Each word and each relationship in the semantic network reflects what is important to social actors as they go about ordering their lives together. The article concludes that obshcheye is so definitively a semantic network that expunging its conceptual heterogeneity and narrowing the multiple logics to encompass one in particular would amount to analytical reductionism and the impoverishment of social analysis. [ABSTRACT FROM AUTHOR]

The relative ineffectiveness of semantics-based multimedia indexing systems on the Web is caused by the semantic knowledge incompleteness and semantic heterogeneity problems. Nevertheless, the need to search multimedia documents with precision on the Web is persistently growing; pressing the demand for effective and efficient indexing strategies. In this article, we present an ontology-based multimedia indexing approach that cooperatively identifies the semantic and taxonomic relations that exist between annotation words that surround multimedia documents on Webpages. In this context, multiple ontologies are jointly employed for indexing each document. We construct inverted indexes in the form of semantic networks where nodes of each network are identified and added based on a majority-voting technique, while edges represent the semantic and taxonomic relations that hold between those nodes. To alleviate the heterogeneity between the resulting networks, we employ ontology merging algorithms to integrate them into consistent networks. We also utilise concept relatedness measures to enrich the networks with semantically-relevant entities that are not recognised by the used ontologies. To validate our proposal, we have developed a prototype system based on the proposed techniques. The produced results using real-world datasets demonstrate an improvement of the effectiveness against state-of-the-art baseline metrics. [ABSTRACT FROM AUTHOR]

Semantic memory encompasses knowledge of specific objects and their diverse associations, but the mechanisms that allow us to retrieve aspects of knowledge required for a given task are poorly understood. The dual hub theory suggests that separate semantic stores represent knowledge of (i) taxonomic categories (in the anterior temporal lobes, ATL) and (ii) thematic associations (in angular gyrus, AG or posterior middle temporal gyrus, pMTG). Alternatively, the controlled semantic cognition (CSC) framework suggests that semantic processing emerges from the flexible interaction of heteromodal semantic representations in ATL with a semantic control network, which includes pMTG as well as prefrontal regions. According to this view, ATL supports patterns of coherent auto-associative retrieval, while semantic control sites respond when ongoing conceptual activation needs to be altered to suit the task or context. These theories make different predictions about the nature of functional dissociations within the semantic network. We review evidence for these claims across multiple methods. First, we show ATL is sensitive to the strength of thematic associations as well as taxonomic relations. Next, we document functional dissociations between AG and pMTG: rather than these regions acting as comparable thematic hubs, AG is allied to the default mode network and supports more 'automatic' retrieval, while pMTG responds when control demands are high. However, the semantic control network, including pMTG, also shows a greater response to events/actions and verbs, supporting the claims of both theories. We propose that tasks tapping event semantics often require greater shaping of conceptual retrieval than comparison tasks, because these elements of our knowledge are inherently flexible, with relevant features depending on the context. In this way, the CSC account might be able to account for findings that suggest both a process and a content distinction within the semantic network. This article is part of the theme issue 'Towards mechanistic models of meaning composition'. [ABSTRACT FROM AUTHOR]

Despite the significant advances noted in semantic segmentation of aerial imagery, a considerable limitation is blocking its adoption in real cases. If we test a segmentation model on a new area that is not included in its initial training set, accuracy will decrease remarkably. This is caused by the domain shift between the new targeted domain and the source domain used to train the model. In this paper, we addressed this challenge and proposed a new algorithm that uses Generative Adversarial Networks (GAN) architecture to minimize the domain shift and increase the ability of the model to work on new targeted domains. The proposed GAN architecture contains two GAN networks. The first GAN network converts the chosen image from the target domain into a semantic label. The second GAN network converts this generated semantic label into an image that belongs to the source domain but conserves the semantic map of the target image. This resulting image will be used by the semantic segmentation model to generate a better semantic label of the first chosen image. Our algorithm is tested on the ISPRS semantic segmentation dataset and improved the global accuracy by a margin up to 24% when passing from Potsdam domain to Vaihingen domain. This margin can be increased by addition of other labeled data from the target domain. To minimize the cost of supervision in the translation process, we proposed a methodology to use these labeled data efficiently. [ABSTRACT FROM AUTHOR]

A virtual enterprise (VE) involves the coordination of collaborative partners in the supply chain network. To seek the market opportunity, the VE initiator and the VE partners have to negotiate with each other to achieve mutual gains as well as ultimate customer satisfaction. Messages sent and received by the counterparts during the negotiation process are considered as negotiation knowledge. However, it is difficult for different companies involved in the VE to recognise the heterogeneous knowledge on their own, especially in an agent-based negotiation system. Thus, semantic interoperability is an important issue in a VE. The objective of this research is to propose an ontology-based knowledge representation approach to provide a semantic interoperable environment to realise automatic negotiations in the VE. The contribution of this article is twofold. On the one hand, an ontology-based knowledge representation scheme is established for VE partners' ontology construction, matching and integration. On the other hand, two agent interaction protocols are developed for the multi-agent system (MAS) to retrieve semantic correspondences and conduct bid negotiations. [ABSTRACT FROM AUTHOR]

As a powerful forgery operation in the image content security area, image inpainting based on deep generative networks can yield visually appealing outputs but often produces ambiguous artifacts, especially in boundary and high semantic areas. To address this issue, a novel end-to-end network with gradient semantics and spatial-smooth attention (GS-SSA) is proposed, which combines a gradient learning network and an image inpainting network. The gradient learning network is meant to properly anticipate the gradient semantics in the hole region and get a complete gradient semantic map. The image inpainting network utilizes the complete gradient semantic map to better repair the missing pixels and obtain the final inpainting results. Moreover, spatial-smooth attention is introduced into the image inpainting network, considering both the spatial structural relations and the surrounding information in the hole region. Experimental results on public datasets show the superiority of the proposed method, especially in large-hole region tasks. • Deep generative network for image inpainting in large-hole region tasks. • Gradient semantics is utilized to acquire the structure and semantic information. • Spatial-smooth considers both the spatial relations and adjacency relations. • Image inpainting of both irregular and regular damaged areas. [ABSTRACT FROM AUTHOR]

Examines the property of semantic correctness of programs written in a shared-memory parallel language. Requirements in parallel programming; Importance of sequential consistency of a multiprocess device to parallel programming; Theorem of semantic equivalence.

Previous research demonstrates that rhetorical figures differentially affect the extent of ad processing. Specifically, tropes (a type of figure) deviate more from expected language use than schemes, with the greater deviation yielding more extensive ad processing. We extend previous research in two ways by focusing on the incongruity differences that exist between schemes and tropes. Study 1 uses syndicated data (Starch readership scores) to test how figures combine to affect the extent of processing. Results show that when figures leverage unique mechanisms (i.e., schemes and tropes), their combination yields incremental processing gains. Alternatively, when figures leverage redundant mechanisms (e.g., multiple tropes), their combination yields no incremental processing. Study 2 is an experiment that tests how figures separate in affecting the focus of ad processing. Results show that schemes generate a generalized focus on the entire ad, including both ad-stylistic and message-related aspects, while tropes generate a more selective focus on message-related aspects. [ABSTRACT FROM AUTHOR]

Very large-scale conversation (VLSC) involves the exchange of thousands of electronic mail (e-mail) messages among hundreds or thousands of people. Usenet newsgroups are good examples (but not the only examples) of online sites where VLSCs take place. To facilitate understanding of the social and semantic structure of VLSCs, two tools from the social sciences--social networks and semantic networks--have been extended for the purposes of interface design. As interface devices, social and semantic networks need to be flexible, layered representations that are useful as a means for summarizing, exploring, and cross-indexing the large volumes of messages that constitute the archives of VLSCs. This paper discusses the design criteria necessary for transforming these social scientific representations into interface devices. The discussion is illustrated with the description of the Conversation Map system, an implemented system for browsing and navigating VLSCs. [ABSTRACT FROM AUTHOR]

The majority of semantic segmentation networks generally employ cross-entropy as a loss function and intersection-over-union (IoU) as the evaluation metric for network performance. Employing IoU as a loss function can solve the mismatch issue between the loss function and the evaluation metric. We propose a Soft IoU training strategy based on mini-batch (mini-batch Soft IoU). Our work has two primary contributions: The first is to extend the IoU loss function to a multi-class segmentation network. The second is to collect various categories of the training samples in every mini-batch, which will ensure that the number of categories equals to the batch size at least. Our method breaks the randomness of the original mini-batch gradient descent (GD) strategy, advancing training samples in the mini-batch much more consistent with the distribution characteristics of the overall data. It solves the instability of IoU loss function. In addition, the experimental results on the PASCAL VOC2012 dataset reveal that our method effectively improves the segmentation accuracy of the network and attains significant improvements beyond state-of-the-art IoU loss function methods. [ABSTRACT FROM AUTHOR]