Search

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]

Knowledge representation is one of the most explored areas in nowadays computer science research. In this paper authors pursue definition and semantics of semantic networks that are defined as part of Semantic Knowledge Base being a hybrid knowledge oriented system. The approach presented in here aims at introducing advanced properties of networks such as cardinality, partitioning or certainty at the same time using simple structure based on two operands and operators. Following paper is an extension of a conference publication that introduced advanced aspects of semantic networks modelling with the use of Association-Oriented Metamodel. The extension includes a discussion related to the formal description of the structure, as well as the description and use of association-oriented design patterns. [ABSTRACT FROM AUTHOR]

Images obtained from an illuminated scene often have their original colour contaminated. Colour constancy is a study considering how to restore them. Substantial progress on colour constancy has been made in recent years due to the development of a convolutional neural network (CNN). In a CNN structure, high-level features contain semantic information while low-level features show local details. If both are taken into account, they would help achieve a more accurate illuminant estimation. However, previous works paid little attention to the latter for lack of frameworks, which can combine those two kinds of features together. Inspired by the pyramid model, a top-down network that successively propagates high-level information to low-level layers is proposed. This network, named top-down semantic aggregation for colour constancy (TDCC), takes full advantage of the multi-scale representations with strong semantics. As a result, objects with intrinsic colours are captured and a better estimation is obtained. Experiments on three benchmark datasets demonstrate that TDCC significantly outperforms stateof- the-art colour constancy methods. [ABSTRACT FROM AUTHOR]

In this article the author offers response to Elliott Oring perception on oppositions, overlaps and ontologies. Topics include considered that Script-based Semantic Theory of Humor proposed as semantic theory around the time than other script-like developments discussed within the Artificial Intelligence community; and realized that human conversations and scenarios take place in isolation of the overall world knowledge.

This paper introduces a method for creating scales of constructs based on word bigram cooccurrences in natural language text. Instead of using a stop-word list to drop less useful words, we use a start-word list that enables computing the cooccurrences of only these "smart words." In this way, we can create scales to measure communication constructs by first listing the key terms in the conceptual definition, then expanding the terms by looking up synonyms in dictionaries such as WordNet. Following this, we compute the cooccurrence network among these words with a sliding window. Next, we extract the first dimension through principal components analysis and identify the words that load most highly on the first factor. For these words, we sum the frequencies, which produces the final index for the construct. This operationalization yields index scales that have high construct validity, which contributes to external validity. Extending the procedures of classic psychometric index construction into the natural language domain avoids the biases of data based on fixed-choice questionnaires. To demonstrate the procedures for construct operationalization, we use a dataset of news stories about the BP Deepwater Horizon Gulf Oil Spill, scaling environmental uncertainty and organizational responses to it including innovation, strategic planning, and changes in organizational structure. [ABSTRACT FROM AUTHOR]

This paper suggests a method for refining a massive amount of collective intelligence data and visualizing it with a multilevel sentiment network in order to understand the relevant information in an intuitive and semantic way. This semantic interpretation method minimizes network learning in the system as a fixed network topology only exists as a guideline to help users understand. Furthermore, it does not need to discover every single node to understand the characteristics of each clustering within the network. After extracting and analyzing the sentiment words from the movie review data, we designed a movie network based on the similarities between the words. The network formed in this way will appear as a multilevel sentiment network visualization after the following three steps: (1) design a heatmap visualization to effectively discover the main emotions on each movie review; (2) create a two-dimensional multidimensional scaling (MDS) map of semantic word data to facilitate semantic understanding of network and then fix the movie network topology on the map; (3) create an asterism graphic with emotions to allow users to easily interpret node groups with similar sentiment words. The research also presents a virtual scenario about how our network visualization can be used as a movie recommendation system. We next evaluated our progress to determine whether it would improve user cognition for multilevel analysis experience compared to the existing network system. Results showed that our method provided improved user experience in terms of cognition. Thus, it is appropriate as an alternative method for semantic understanding. [ABSTRACT FROM AUTHOR]

Introduction. Truth and falsity play a dominant role in contemporary work on information, misinformation, and disinformation. Discussions of these notions typically include remarks about their truth-values, or even explicit arguments in favour of their having fixed truth-values. Method. I use Chalmers’s framework for identifying and potentially solving verbal disputes to analyse the discussions regarding information and truth. Analysis. Philosophical analyses of the discussions of information and its relation to truth, as they play out between Luciano Floridi and Don Fallis (among others), are carried out in the paper. Results. I find that these discussions are in fact a verbal dispute. Further, the dominant focus on truth and falsity in the work on information, misinformation, and disinformation, within philosophy of information, gives rise to a Floridian dilemma: a dilemma in which Floridi’s definition of semantic information, as inherently truthful, turns out to encompass certain varieties of misinformation and disinformation while excluding others. Conclusion. I recommend that information must be defined as semantic content in general without reference to truth, i.e. as truth-neutral, such that all varieties of misinformation and disinformation can be kinds of information. [ABSTRACT FROM AUTHOR]

A defining characteristic of Alzheimer's disease is difficulty in retrieving semantic memories, or memories encoding facts and knowledge. While it has been suggested that this impairment is caused by a degradation of the semantic store, the precise ways in which the semantic store is degraded are not well understood. Using a longitudinal corpus of semantic fluency data (listing of items in a category), we derive semantic network representations of patients with Alzheimer's disease and of healthy controls. We contrast our network-based approach with analyzing fluency data with the standard method of counting the total number of items and perseverations in fluency data. We find that the networks of Alzheimer's patients are more connected and that those connections are more randomly distributed than the connections in networks of healthy individuals. These results suggest that the semantic memory impairment of Alzheimer's patients can be modeled through the inclusion of spurious associations between unrelated concepts in the semantic store. We also find that information from our network analysis of fluency data improves prediction of patient diagnosis compared to traditional measures of the semantic fluency task. [ABSTRACT FROM AUTHOR]

Sparse Representation-based Classifier (SRC) and Dictionary Learning (DL), have significantly impacted greatly on the classification performance of image recognition in recent times. In video semantic analysis, the locality structure of video semantic data containing more discriminative information is very essential for classification. However, this has not been fully considered by the current sparse representation-based approaches. Furthermore, similar coding outcomes are not being realized from video features with the same video category. To handle these issues, we propose a novel DL method, called Group Sparsity Locality-Sensitive Dictionary Learning (GSLSDL) for video semantic analysis. In the proposed GSLSDL, a discriminant loss function for the video category based on group sparse coding of sparse coefficients, is introduced into the structure of the Locality-Sensitive Dictionary Learning (LSDL) method. After solving the optimized dictionary, the sparse coefficients for the testing video feature samples are obtained. The classification result for video semantic is then realized by minimizing the error between the original and reconstructed samples. The experiment results show that, the proposed GSLSDL significantly improves the performance of video semantic detection compared with the competing methods, and robust in various diverse environments of video. [ABSTRACT FROM AUTHOR]

Abstract Remote Sensing technologies allow to map biophysical, biochemical, and earth surface parameters of the land surface. Of especial interest for various applications in environmental and urban sciences is the combination of spectral and 3D elevation information. However, those two data streams are provided separately by different instruments, namely airborne laser scanner (ALS) for elevation and a hyperspectral imager (HSI) for high spectral resolution data. The fusion of ALS and HSI data can thus lead to a single data entity consistently featuring rich structural and spectral information. In this study, we present the application of fusing the first pulse return information from ALS data at a sub-decimeter spatial resolution with the lower-spatial resolution hyperspectral information available from the HSI into a hyperspectral point cloud (HSPC). During the processing, a plausible hyperspectral spectrum is assigned to every first-return ALS point. We show that the complementary implementation of spectral and 3D information at the point-cloud scale improves object-based classification and information extraction schemes. This improvements have great potential for numerous land-cover mapping and environmental applications. [ABSTRACT FROM AUTHOR]

Declarative memory enables cognitive agents to effectively store and retrieve factual memory in real-time. Increasing the capacity of a real-time agent's declarative memory increases an agent's ability to interact intelligently with its environment but requires a scalable retrieval system. This work represents an extension of the Accelerated Declarative Memory (ADM) system, referred to as Hardware Accelerated Declarative Memory (HADM), to execute retrievals on a GPU. HADM also presents improvements over ADM's CPU execution and considers critical behavior for indefinitely running declarative memories. The negative effects of a constant maximum associative strength are considered, and mitigating solutions are proposed. HADM utilizes a GPU to process the entire semantic network in parallel during retrievals, yielding significantly faster declarative retrievals. The resulting GPU-accelerated retrievals show an average speedup of approximately 70 times over the previous Service Oriented Architecture Declarative Memory (soaDM) implementation and an average speedup of approximately 5 times over ADM. HADM is the first GPU-accelerated declarative memory system in existence. [ABSTRACT FROM AUTHOR]

Objective: Positive symptoms, such as delusion and hallucination, commonly include negative emotional content in schizophrenia. We investigated the neural basis implicated during the processing of strong negative emotional words in patients with schizophrenia. Methods: In our study, 35 patients with schizophrenia and 19 healthy controls were recruited, and the participants were asked to passively view the words that contained swearing and neutral content during functional magnetic resonance imaging. Results: Patients with schizophrenia, compared to healthy controls, showed hypoactivation to the swear and neutral words stimuli in the left inferior frontal gyrus, left middle frontal gyrus, and left angular/supramarginal gyrus. More specifically, patients with remitted schizophrenia were found to have greater activation to the stimuli in the left middle/inferior frontal gyrus than patients with active schizophrenia. Furthermore, in the analysis of regions of interests, the left inferior and middle frontal gyrus activity was related to the severity of positive symptoms, including delusion and suspiciousness. Conclusion: Our results suggest that patients with schizophrenia have difficulty in semantic processing and inhibitory control of swear words, and these abnormalities may be connected with the severity of positive symptoms. [ABSTRACT FROM AUTHOR]

Document Standard Operating Procedures (SOP) can manage business processes and employee performance in an organization. This study aims to develop a system that can publish SOP documents automatically distributed to employees. In this study, an analysis of the relationship between SOP documents and employees is carried out so that it can be directly allocated to employees according to the position. Knowledge analysis of the relationship between SOP documents and employees is done using semantic network analysis. Semantic networks are used to analyze components of knowledge and the relationship between SOP documents and employees. The results of the report of the elements of knowledge and the relationship between SOP documents found 25 SOP documents were consisting of 6 types of central nodes with 156 child nodes and had 7 types of relations containing 207 relations. SOP knowledge management system is connected to the personnel information system (SIPEG) so that it makes it easier for users to find, accommodate, and manage the knowledge contained in SOP documents. System implementation is done using PHP programming language with CodeIgniter framework, Rest API, and MySQL database. [ABSTRACT FROM AUTHOR]

POOL—A Semantic Model for Approximate Reasoning—is based on the theory of linguistic fuzzy relations (LFR) and linguistic similarity relations (LSR). The proposed model combines the advantages of linguistic representation and numerical computation. Hence, it can facilitate knowledge representation and approximate reasoning for object/concept relation modeling especially in an environment containing imprecision, ambiguity, and/or uncertainty, which is often the case faced by knowledge engineers in building decision support systems. [ABSTRACT FROM AUTHOR]

The concept of an electronic marketplace is introduced whereby consumers may browse through a structured database giving product offerings and local supply outlets. We consider the nature of software to facilitate the shopping process developing a concept of similarity between generic products using a semantic net structure. The search is guided by specified user preferences. A prototype implementation, using Prolog, is described. [ABSTRACT FROM AUTHOR]

Transfer learning, which focuses on finding a favorable representation for instances of different domains based on auxiliary data, can mitigate the divergence between domains through knowledge transfer. Recently, increasing efforts on transfer learning have employed deep neural networks (DNN) to learn more robust and higher level feature representations to better tackle cross-media disparities. However, only a few articles consider the correction and semantic matching between multi-layer heterogeneous domain networks. In this article, we propose a deep semantic mapping model for heterogeneous multimedia transfer learning (DHTL) using co-occurrence data. More specifically, we integrate the DNN with canonical correlation analysis (CCA) to derive a deep correlation subspace as the joint semantic representation for associating data across different domains. In the proposed DHTL, a multi-layer correlation matching network across domains is constructed, in which the CCA is combined to bridge each pair of domain-specific hidden layers. To train the network, a joint objective function is defined and the optimization processes are presented. When the deep semantic representation is achieved, the shared features of the source domain are transferred for task learning in the target domain. Extensive experiments for three multimedia recognition applications demonstrate that the proposed DHTL can effectively find deep semantic representations for heterogeneous domains, and it is superior to the several existing state-of-the-art methods for deep transfer learning. [ABSTRACT FROM AUTHOR]

Highlights • Extending the natural language to the design intent representation for collaborative product design. • Supporting designers to share and exchange the design intents in knowledge-based collaborative CAD. • Combining collaborative design, ontology, semantic network and natural language generation together. • The ability in recognizing and repairing features, fixing design history and picking out features. Abstract Modern manufacturing firms are more inclining to promote the product quality, save costs and reduce times of product design by both collaborative designing and model reuse. If CAD components constructed collaboratively have information representing their developers' design intents embedded in the model, people's understanding over the product should be improved and the product model should be best reused. Until now, capturing, recording and presenting design intents still remains a challenge. It has been shown by empirical studies that textual summarisations can lead to improved decision making. In this paper, we propose an approach to generation the natural language description about design intents of collaboratively developed product. The approach brings together techniques from different areas of collaborative designing, ontology and semantic network, and natural language generation. The language generation process is guided by an information model we established to give a structured description about design intents of collaboratively products. In order to record information related to the design intents, we build a common CAD model ontology and then generate a semantic network to describe dependencies, component structures and design history which are components of the design intent information model. The techniques of natural language generation, namely discourse planning and sentence planning, are adopted for the eventual linguistic generation of design intents. Finally, we use several case studies to prove the advantages of natural language in helping people better understanding the design intents. [ABSTRACT FROM AUTHOR]

Research on eudaimonic media has so far predominantly focused on audiovisual offerings: movies or YouTube video clips. However, much meaningful and inspiring content nowadays is uploaded on social media in so-called memes. Three exploratory studies therefore investigated the occurrence, content, and effects of inspiring and meaningful memes in social media: The hashtags of eudaimonic memes were analyzed in semantic networks (study 1), a content analysis was conducted to typologize eudaimonic themes addressed in memes (study 2), and an online survey investigated the effects of eudaimonic memes on users (study 3). The results suggest that previously defined topics of inspiration and meaningfulness are also common among hashtags and memes and lead to the same beneficial effects for their consumers. The discussion aims at advancing the theory of mediated eudaimonia and understanding its relevance for well-being in the daily lives of social media users. [ABSTRACT FROM AUTHOR]