Adaptive Graph Convolutional Adjacency Matrix Network for Video Summarization

Video summarization aims to select key frames or key shots to create summaries for fast retrieval,compression,and efficient browsing of videos.Graph neural networks efficiently capture information about graph nodes and their neighbors,but ignore the dynamic dependencies between nodes.To address this challenge,we propose an innovative Adaptive Graph Convolutional Adjacency Matrix Network(TAMGCN),leveraging the attention mechanism to dynamically adjust dependencies between graph nodes.Specifically,we first segment shots and extract features of each frame,then compute the representative features of each shot.Subsequently,we utilize the attention mechanism to dynamically adjust the adjacency matrix of the graph convolutional network to better capture the dynamic dependencies between graph nodes.Finally,we fuse temporal features extracted by Bi-directional Long Short-Term Memory network with structural features extracted by the graph convolutional network to generate high-quality summaries.Extensive experiments are conducted on two benchmark datasets,TVSum and SumMe,yielding F1-scores of 60.8%and 53.2%,respectively.Experimental results demonstrate that our method outperforms most state-of-the-art video summarization techniques.

Enhanced Topic-Aware Summarization Using Statistical Graph Neural Networks

The rapid expansion of online content and big data has precipitated an urgent need for efficient summarization techniques to swiftly comprehend vast textual documents without compromising their original integrity.Current approaches in Extractive Text Summarization(ETS)leverage the modeling of inter-sentence relationships,a task of paramount importance in producing coherent summaries.This study introduces an innovative model that integrates Graph Attention Networks(GATs)with Transformer-based Bidirectional Encoder Representa-tions from Transformers(BERT)and Latent Dirichlet Allocation(LDA),further enhanced by Term Frequency-Inverse Document Frequency(TF-IDF)values,to improve sentence selection by capturing comprehensive topical information.Our approach constructs a graph with nodes representing sentences,words,and topics,thereby elevating the interconnectivity and enabling a more refined understanding of text structures.This model is stretched to Multi-Document Summarization(MDS)from Single-Document Summarization,offering significant improvements over existing models such as THGS-GMM and Topic-GraphSum,as demonstrated by empirical evaluations on benchmark news datasets like Cable News Network(CNN)/Daily Mail(DM)and Multi-News.The results consistently demonstrate superior performance,showcasing the model’s robustness in handling complex summarization tasks across single and multi-document contexts.This research not only advances the integration of BERT and LDA within a GATs but also emphasizes our model’s capacity to effectively manage global information and adapt to diverse summarization challenges.

Generation of scale-free knowledge network with local world mechanism

In order to simulate the real growing process, a new type of knowledge network growth mechanism based on local world connectivity is constructed. By the mean-field method, theoretical prediction of the degree distribution of the knowledge network is given, which is verified by Matlab simulations. When the new added node＇s local world size is very small, the degree distribution of the knowledge network approximately has the property of scale-free. When the new added node＇s local world size is not very small, the degree distribution transforms from pure power-law to the power-law with an exponential tailing. And the scale-free index increases as the number of new added edges decreases and the tunable parameters increase. Finally, comparisons of some knowledge indices in knowledge networks generated by the local world mechanism and the global mechanism are given. In the long run, compared with the global mechanism, the local world mechanism leads the average knowledge levels to slower growth and brings homogenous phenomena.

Topic-Aware Abstractive Summarization Based on Heterogeneous Graph Attention Networks for Chinese Complaint Reports

Automatic text summarization(ATS)plays a significant role in Natural Language Processing(NLP).Abstractive summarization produces summaries by identifying and compressing the most important information in a document.However,there are only relatively several comprehensively evaluated abstractive summarization models that work well for specific types of reports due to their unstructured and oral language text characteristics.In particular,Chinese complaint reports,generated by urban complainers and collected by government employees,describe existing resident problems in daily life.Meanwhile,the reflected problems are required to respond speedily.Therefore,automatic summarization tasks for these reports have been developed.However,similar to traditional summarization models,the generated summaries still exist problems of informativeness and conciseness.To address these issues and generate suitably informative and less redundant summaries,a topic-based abstractive summarization method is proposed to obtain global and local features.Additionally,a heterogeneous graph of the original document is constructed using word-level and topic-level features.Experiments and analyses on public review datasets(Yelp and Amazon)and our constructed dataset(Chinese complaint reports)show that the proposed framework effectively improves the performance of the abstractive summarization model for Chinese complaint reports.

Abstractive Arabic Text Summarization Using Hyperparameter Tuned Denoising Deep Neural Network

ive Arabic Text Summarization using Hyperparameter Tuned Denoising Deep Neural Network(AATS-HTDDNN)technique.The presented AATS-HTDDNN technique aims to generate summaries of Arabic text.In the presented AATS-HTDDNN technique,the DDNN model is utilized to generate the summary.This study exploits the Chameleon Swarm Optimization(CSO)algorithm to fine-tune the hyperparameters relevant to the DDNN model since it considerably affects the summarization efficiency.This phase shows the novelty of the current study.To validate the enhanced summarization performance of the proposed AATS-HTDDNN model,a comprehensive experimental analysis was conducted.The comparison study outcomes confirmed the better performance of the AATS-HTDDNN model over other approaches.

New Knowledge Network Evaluation Method for Design Rationale Management

Current design rationale （DR） systems have not demonstrated the value of the approach in practice since little attention is put to the evaluation method of DR knowledge. To systematize knowledge management process for future computer-aided DR applications, a prerequisite is to provide the measure for the DR knowledge. In this paper, a new knowledge network evaluation method for DR management is presented. The method characterizes the DR knowledge value from four perspectives, namely, the design rationale structure scale, association knowledge and reasoning ability, degree of design justification support and degree of knowledge representation conciseness. The DR knowledge comprehensive value is also measured by the proposed method. To validate the proposed method, different style of DR knowledge network and the performance of the proposed measure are discussed. The evaluation method has been applied in two realistic design cases and compared with the structural measures. The research proposes the DR knowledge evaluation method which can provide object metric and selection basis for the DR knowledge reuse during the product design process. In addition, the method is proved to be more effective guidance and support for the application and management of DR knowledge.

Modeling and Robustness of Knowledge Network in Supply Chain

The growth and evolution of the knowledge network in supply chain can be characterized by dynamic growth clustering and non-homogeneous degree distribution.The networks with the above characteristics are also known as scale-free networks.In this paper,the knowledge network model in supply chain is established,in which the preferential attachment mechanism based on the node strength is adopted to simulate the growth and evolution of the network.The nodes in the network have a certain preference in the choice of a knowledge partner.On the basis of the network model,the robustness of the three network models based on different preferential attachment strategies is investigated.The robustness is also referred to as tolerances when the nodes are subjected to random destruction and malicious damage.The simulation results of this study show that the improved network has higher connectivity and stability.

Development of an Ontology-Based Knowledge Network by Interconnecting Soil/Water Concepts/Properties, Derived from Standards Methods and Published Scientific References Outlining Infiltration/Percolation Process of Contaminated Water

The present work deals with the development of an Ontology-Based Knowledge Network of soil/water physicochemical & biological properties (soil/water concepts), derived from ASTM Standard Methods (ASTMi,n) and relevant scientific/applicable references (published papers—PPi,n) to fill up/bridge the gap of the information science between cited Standards and infiltration discipline conceptual vocabulary providing accordingly a dedicated/internal Knowledge Base (KB). This attempt constitutes an innovative approach, since it is based on externalizing domain knowledge in the form of Ontology-Based Knowledge Networks, incorporating standardized methodology in soil engineering. The ontology soil/water concepts (semantics) of the developed network correspond to soil/water physicochemical & biological properties, classified in seven different generations that are distinguished/located in infiltration/percolation process of contaminated water through soil porous media. The interconnections with arcs between corresponding concepts/properties among the consecutive generations are defined by the relationship of dependent and independent variables. All these interconnections are documented according to the below three ways: 1) dependent and independent variables interconnected by using the logical operator “depends on” quoting existent explicit functions and equations;2) dependent and independent variables interconnected by using the logical operator “depends on” quoting produced implicit functions, according to Rayleigh’s method of indices;3) dependent and independent variables interconnected by using the logical operator “related to” based on a logical dependence among the examined nodes-concepts-variables. The aforementioned approach provides significant advantages to semantic web developers and web users by means of prompt knowledge navigation, tracking, retrieval and usage.

Structures and dynamics of scientific knowledge networks:An empirical analysis based on a co-word network

Co-word networks are constructed with author-provided keywords in academic publications and their relations of co-occurrence.As special form of scientific knowledge networks,they represent the cognitive structure of scientific literature.This paper analyzes the complex structure of a co-word network based on 8,190 author-provided keywords extracted from 3,651 papers in five Chinese core journals in the field of management science.Small-world and scale-free phenomena are found in this network.A large-scale co-word network graph,which consists of one major giant component and many small isolated components,has been generated with the GUESS software.The dynamic growth of keywords and keyword co-occurrence relationships are described with four new informetrics measures.The results indicate that existing concepts always serve as the intellectual base of new ideas as represented by keywords.

Spread spectrum of knowledge network based on information integration environment

The concept and the characteristic of spread spectrum technology of knowledge network based on information integration environment are given at first. And the basic principle of the wireless network of knowledge and information is recommended. Then a project alternatives using the spread spectrum technology and wireless network is given to build a across-the abroad knowledge network system.

Structure, Evolution and Hot Spots of Cooperation Innovation Knowledge Network

Based on the statistics analysis of cooperation innovation articles, the paper analyses the distributing characteristics of cooperation innovation research papers. The distribution, structure and evolution of cooperation innovation research are studied through the social network and co-word method, combined with the information visualization technology. This study is based on the papers in the SSCI database during 2000-2014. The analysis and statistics of papers reveal distribution characteristics of the subject, time, institutions, countries and areas. The statistics shows that the research perspective of cooperation innovation from business and economics. The time distribution presents that research of cooperation innovation attracts enormous attention and the research achievement is increasing. From the distribution of organizations, countries and areas, North America and Europe countries maintain the leading position in this research field. But institutions and universities of China and other Asian countries or areas are witnessed their outstanding achievements in cooperation innovation research. Combining the co-word network analysis method, the paper studies the structure and content of cooperation innovation knowledge network in five stages and makes the net spectrum, visually showing the hot spots at various periods. The top highly cited papers in all five stages are reviewed simultaneously;their research hot spots and evolution process are concluded. The study shows 11 subjects including strategic alliances, social networks, R&D cooperation, R&D, technology transfer, alliances, knowledge management, social capital, entrepreneurship, trust, biotechnology are always the central issues in last 15 years. The focus of research is from the relationship between alliance and technology innovation to networking and knowledge innovation and then to open innovation.

A Semantic Supervision Method for Abstractive Summarization

In recent years,many text summarization models based on pretraining methods have achieved very good results.However,in these text summarization models,semantic deviations are easy to occur between the original input representation and the representation that passed multi-layer encoder,which may result in inconsistencies between the generated summary and the source text content.The Bidirectional Encoder Representations from Transformers(BERT)improves the performance of many tasks in Natural Language Processing(NLP).Although BERT has a strong capability to encode context,it lacks the fine-grained semantic representation.To solve these two problems,we proposed a semantic supervision method based on Capsule Network.Firstly,we extracted the fine-grained semantic representation of the input and encoded result in BERT by Capsule Network.Secondly,we used the fine-grained semantic representation of the input to supervise the fine-grained semantic representation of the encoded result.Then we evaluated our model on a popular Chinese social media dataset(LCSTS),and the result showed that our model achieved higher ROUGE scores(including R-1,R-2),and our model outperformed baseline systems.Finally,we conducted a comparative study on the stability of the model,and the experimental results showed that our model was more stable.

An Intelligent Tree Extractive Text Summarization Deep Learning

In recent research,deep learning algorithms have presented effective representation learning models for natural languages.The deep learningbased models create better data representation than classical models.They are capable of automated extraction of distributed representation of texts.In this research,we introduce a new tree Extractive text summarization that is characterized by fitting the text structure representation in knowledge base training module,and also addresses memory issues that were not addresses before.The proposed model employs a tree structured mechanism to generate the phrase and text embedding.The proposed architecture mimics the tree configuration of the text-texts and provide better feature representation.It also incorporates an attention mechanism that offers an additional information source to conduct better summary extraction.The novel model addresses text summarization as a classification process,where the model calculates the probabilities of phrase and text-summary association.The model classification is divided into multiple features recognition such as information entropy,significance,redundancy and position.The model was assessed on two datasets,on the Multi-Doc Composition Query(MCQ)and Dual Attention Composition dataset(DAC)dataset.The experimental results prove that our proposed model has better summarization precision vs.other models by a considerable margin.

A new evolutional model for institutional field knowledge flow network

Purpose:This paper aims to address the limitations in existing research on the evolution of knowledge flow networks by proposing a meso-level institutional field knowledge flow network evolution model(IKM).The purpose is to simulate the construction process of a knowledge flow network using knowledge organizations as units and to investigate its effectiveness in replicating institutional field knowledge flow networks.Design/Methodology/Approach:The IKM model enhances the preferential attachment and growth observed in scale-free BA networks,while incorporating three adjustment parameters to simulate the selection of connection targets and the types of nodes involved in the network evolution process Using the PageRank algorithm to calculate the significance of nodes within the knowledge flow network.To compare its performance,the BA and DMS models are also employed for simulating the network.Pearson coefficient analysis is conducted on the simulated networks generated by the IKM,BA and DMS models,as well as on the actual network.Findings:The research findings demonstrate that the IKM model outperforms the BA and DMS models in replicating the institutional field knowledge flow network.It provides comprehensive insights into the evolution mechanism of knowledge flow networks in the scientific research realm.The model also exhibits potential applicability to other knowledge networks that involve knowledge organizations as node units.Research Limitations:This study has some limitations.Firstly,it primarily focuses on the evolution of knowledge flow networks within the field of physics,neglecting other fields.Additionally,the analysis is based on a specific set of data,which may limit the generalizability of the findings.Future research could address these limitations by exploring knowledge flow networks in diverse fields and utilizing broader datasets.Practical Implications:The proposed IKM model offers practical implications for the construction and analysis of knowledge flow networks within institutions.It provides a valuable tool for understanding and managing knowledge exchange between knowledge organizations.The model can aid in optimizing knowledge flow and enhancing collaboration within organizations.Originality/value:This research highlights the significance of meso-level studies in understanding knowledge organization and its impact on knowledge flow networks.The IKM model demonstrates its effectiveness in replicating institutional field knowledge flow networks and offers practical implications for knowledge management in institutions.Moreover,the model has the potential to be applied to other knowledge networks,which are formed by knowledge organizations as node units.

GATiT:An Intelligent Diagnosis Model Based on Graph Attention Network Incorporating Text Representation in Knowledge Reasoning

The growing prevalence of knowledge reasoning using knowledge graphs(KGs)has substantially improved the accuracy and efficiency of intelligent medical diagnosis.However,current models primarily integrate electronic medical records(EMRs)and KGs into the knowledge reasoning process,ignoring the differing significance of various types of knowledge in EMRs and the diverse data types present in the text.To better integrate EMR text information,we propose a novel intelligent diagnostic model named the Graph ATtention network incorporating Text representation in knowledge reasoning(GATiT),which comprises text representation,subgraph construction,knowledge reasoning,and diagnostic classification.In the text representation process,GATiT uses a pre-trained model to obtain text representations of the EMRs and additionally enhances embeddings by including chief complaint information and numerical information in the input.In the subgraph construction process,GATiT constructs text subgraphs and disease subgraphs from the KG,utilizing EMR text and the disease to be diagnosed.To differentiate the varying importance of nodes within the subgraphs features such as node categories,relevance scores,and other relevant factors are introduced into the text subgraph.Themessage-passing strategy and attention weight calculation of the graph attention network are adjusted to learn these features in the knowledge reasoning process.Finally,in the diagnostic classification process,the interactive attention-based fusion method integrates the results of knowledge reasoning with text representations to produce the final diagnosis results.Experimental results on multi-label and single-label EMR datasets demonstrate the model’s superiority over several state-of-theart methods.

Automated knowledge graphs for complex systems (AutoGraCS): Applications to management of bridge networks

With the ability to harness the power of big data,the digital twin(DT)technology has been increasingly applied to the modeling and management of structures and infrastructure systems,such as buildings,bridges,and power distribution systems.Supporting these applications,an important family of methods are based on graphs.For DT applications in modeling and managing smart cities,large-scale knowledge graphs(KGs)are necessary to represent the complex interdependencies and model the urban infrastructure as a system of systems.To this end,this paper develops a conceptual framework:Automated knowledge Graphs for Complex Systems(AutoGraCS).In contrast to existing KGs developed for DTs,AutoGraCS can support KGs to account for interdependencies and statistical correlations across complex systems.The established KGs from AutoGraCS can then be easily turned into Bayesian networks for probabilistic modeling,Bayesian analysis,and adaptive decision supports.Besides,AutoGraCS provides flexibility in support of users’need to implement the ontology and rules when constructing the KG.With the user-defined ontology and rules,AutoGraCS can automatically generate a KG to represent a complex system consisting of multiple systems.The bridge network in Miami-Dade County,FL is used as an illustrative example to generate a KG that integrates multiple layers of data from the bridge network,traffic monitoring facilities,and flood water watch stations.

AUTOMATIC PATENT DOCUMFNT SUMMARIZATION FOR COLLABORATIVE KNOWLEDGE SYSTEMS AND SERVICES

Engineering and research teams often develop new products and technologies by referring to inventions described in patent databases. Efficient patent analysis builds R＆D knowledge, reduces new product development time, increases market success, and reduces potential patent infringement. Thus, it is beneficial to automatically and systematically extract information from patent documents in order to improve knowledge sharing and collaboration among R＆D team members. In this research, patents are summarized using a combined ontology based and TF-IDF concept clustering approach. The ontology captures the general knowledge and core meaning of patents in a given domain. Then, the proposed methodology extracts, clusters, and integrates the content of a patent to derive a summary and a cluster tree diagram of key terms. Patents from the International Patent Classification （IPC） codes B25C, B25D, B25F （categories for power hand tools） and B24B, C09G and H011 （categories for chemical mechanical polishing） are used as case studies to evaluate the compression ratio, retention ratio, and classification accuracy of the summarization results. The evaluation uses statistics to represent the summary generation and its compression ratio, the ontology based keyword extraction retention ratio, and the summary classification accuracy. The results show that the ontology based approach yields about the same compression ratio as previous non-ontology based research but yields on average an 11% improvement for the retention ratio and a 14% improvement for classification accuracy.

Video summarization with a graph convolutional attention network

Video summarization has established itself as a fundamental technique for generating compact and concise video, which alleviates managing and browsing large-scale video data. Existing methods fail to fully consider the local and global relations among frames of video, leading to a deteriorated summarization performance. To address the above problem, we propose a graph convolutional attention network(GCAN) for video summarization. GCAN consists of two parts, embedding learning and context fusion, where embedding learning includes the temporal branch and graph branch. In particular, GCAN uses dilated temporal convolution to model local cues and temporal self-attention to exploit global cues for video frames. It learns graph embedding via a multi-layer graph convolutional network to reveal the intrinsic structure of frame samples. The context fusion part combines the output streams from the temporal branch and graph branch to create the context-aware representation of frames, on which the importance scores are evaluated for selecting representative frames to generate video summary. Experiments are carried out on two benchmark databases, Sum Me and TVSum, showing that the proposed GCAN approach enjoys superior performance compared to several state-of-the-art alternatives in three evaluation settings.

Semantic web-based networked manufacturing knowledge retrieval system

To deal with a lack of semantic interoperability of traditional knowledge retrieval approaches, a semantic-based networked manufacturing （NM） knowledge retrieval architecture is proposed, which offers a series of tools for supporting the sharing of knowledge and promoting NM collaboration. A 5-tuple based semantic information retrieval model is proposed, which includes the interoperation on the semantic layer, and a test process is given for this model. The recall ratio and the precision ratio of manufacturing knowledge retrieval are proved to be greatly improved by evaluation. Thus, a practical and reliable approach based on the semantic web is provided for solving the correlated concrete problems in regional networked manufacturing.

Personal summarization from profile networks

Personal profile information on social media like LinkedIn.com and Facebook.com is at the core of many inter- esting applications, such as talent recommendation and con- textual advertising. However, personal profiles usually lack consistent organization confronted with the large amount of available information. Therefore, it is always a challenge for people to quickly find desired information from them. In this paper, we address the task of personal profile summarization by leveraging both textual information and social connection information in social networks from both unsupervised and supervised learning paradigms. Here, using social connec- tion information is motivated by the intuition that people with similar academic, business or social background （e.g., co- major, co-university, and co-corporation） tend to have similar experiences and should have similar summaries. For unsu- pervised learning, we propose a collective ranking approach, called SocialRank, to combine textual information in an in- dividual profile and social context information from relevant profiles in generating a personal profile summary. For super- vised learning, we propose a collective factor graph model, called CoFG, to summarize personal profiles with local tex- tual attribute functions and social connection factors. Exten- sive evaluation on a large dataset from LinkedIn.com demon- strates the usefulness of social connection information in per- sonal profile summarization and the effectiveness of our pro- posed unsupervised and supervised learning approaches.