Search

Building a human-computer conversational system that can communicate with humans is a research hotspot in the field of artificial intelligence. Traditional dialogue systems tend to produce irrelevant and non-information responses, which reduce people’s interest in engaging in a conversation. This often leads to boring conversations. To alleviate this problem, many researchers use external knowledge to assist conversation generation. The accuracy of knowledge selection is the prerequisite to ensure the quality of knowledge conversation. This approach has worked positively to a certain extent, but generally only searches knowledge information based on entity words themselves, without considering the specific conversation context. Therefore, if irrelevant knowledge is retrieved, the quality of conversation generation will be reduced. Motivated by this, we propose a novel neural knowledge-based conversation generation model, named Siamese Network based Posterior Knowledge Selection Model for Knowledge Driven Conversation Generation (SPK-CG) . We have designed a novel knowledge selection mechanism to obtain knowledge information that is highly relevant to the context of the conversation. Specifically, the posterior knowledge distribution is used as a soft label to make the prior distribution consistent with the posterior distribution in the training process. At the same time, in order to narrow the gap between prior and posterior distributions and improve the accuracy of knowledge selection, we leverage siamese network and design multi-granularity matching module for knowledge selection. Compared with previous knowledge-based models, our method can select more appropriate knowledge and use the selected knowledge to generate responses that are more relevant to the conversation context. Extensive automatic and human evaluations demonstrate that our model has advantages over previous baselines.

In the era of social networks, the rapid growth of data mining in information retrieval and natural language processing makes automatic text summarization necessary. Currently, pretrained word embedding and sequence to sequence models can be effectively adapted in social network summarization to extract significant information with strong encoding capability. However, how to tackle the long text dependence and utilize the latent topic mapping has become an increasingly crucial challenge for these models. In this article, we propose a topic-aware extractive and abstractive summarization model named T-BERTSum, based on Bidirectional Encoder Representations from Transformers (BERTs). This is an improvement over previous models, in which the proposed approach can simultaneously infer topics and generate summarization from social texts. First, the encoded latent topic representation, through the neural topic model (NTM), is matched with the embedded representation of BERT, to guide the generation with the topic. Second, the long-term dependencies are learned through the transformer network to jointly explore topic inference and text summarization in an end-to-end manner. Third, the long short-term memory (LSTM) network layers are stacked on the extractive model to capture sequence timing information, and the effective information is further filtered on the abstractive model through a gated network. In addition, a two-stage extractive-abstractive model is constructed to share the information. Compared with the previous work, the proposed model T-BERTSum focuses on pretrained external knowledge and topic mining to capture more accurate contextual representations. Experimental results on the CNN/Daily mail and XSum datasets demonstrate that our proposed model achieves new state-of-the-art results while generating consistent topics compared with the most advanced method.

Currently, building an end-to-end dialogue system for multi-domain task-oriented dialogue has some enormous challenges. Dialogue systems must obtain entire dialogue states from all relevant domains in order to respond correctly. However, multiple domains are involved in the actual dialogue process, which increases the difficulty of obtaining the dialogue state. In addition, dialogue systems process diverse information such as dialogue history, dialogue state, dialogue act, and database across domains, resulting in natural responses. These complex dialogue information brings greater difficulties to response generation. In this paper, we propose a novel unified dialogue framework for multi-domain dialogue in task-oriented dialog, including three modules: Encoder, Dialogue State Tracker and Multiple Decoders. First, encoder module encode all text input into continuous representations. Secondly, we train the dialogue state tracker module with a stacked-attention architecture.It learns information from slot-atten structure and domain-atten structure to track dialogue state. Then, multiple decoders module consists of act decoder structure and response decoder structure. It combines information from different textual inputs while modeling dialogue act. Finally, we jointly train all the above modules to generate system responses. We conducted extensive experiments on the dataset MultiWOZ. The experimental results show that our model achieves state-of-the-art results on evaluation metrics compared to models from previous work.

Clustering is an important research direction in data mining. However, there is no one clustering algorithm that can be applied efficiently in all situation. Clustering ensemble is the best way to solve the above-mentioned problems. It combines the results of multiple clustering algorithms, and the final result is significantly better than a single clustering algorithm. Although there is a lot of constraint information, the existing clustering ensemble algorithm does not utilize it. This paper uses constraint information in consensus function and proposes a Semi-supervised Selective Clustering Ensemble based on Chameleon (SSCEC) and Semi-supervised Selective Clustering Ensemble based on Ncut (SSCEN) to solve the above problem. SSCEC uses the chameleon algorithm as consensus function, and processes constraint information in subgraph partition and subgraph combining. SSCEN uses the Normalized cut algorithm as consensus function, and processes constraint information in the process of graph dichotomy. The experiment results show that our proposed two semi-supervised member selection clustering ensemble algorithms are better than other semi-supervised algorithms.

Social media has recently become one of the most used media in the world. This has resulted in a great hotbed for the growth of rumors, as anyone can spread knowledge and opinions without confirmation. Previous works on rumor detection focused on hand-extracted features and spent less effort on text representation. In this research, a novel method for rumor detection on social media is proposed, which integrates entity recognition, sentence reconfiguration and ordinary differential equation network under a unified framework called ESODE. An entity recognition method to enhance the semantic understanding of rumor texts is used. Then, a sentence reconfiguration to improve the frequency of important words is designed. The complete feature map is established by further collecting statistical features from three aspects: linguistic features on the content of rumors, characteristics of users involved in rumor propagating, and propagation network structures. Finally, the ordinary differential equation network (ODEnet) is applied to detect rumors. Experimental results on datasets from Twitter and Weibo show that the proposed method achieves better performance than previous ones.

The typical aim of user matching is to detect the same individuals cross different social networks. The existing efforts in this field usually focus on the users' attributes and network embedding, but these methods often ignore the closeness between the users and their friends. To this end, we present a friend closeness based user matching algorithm (FCUM). It is a semi-supervised and end-to-end cross networks user matching algorithm. Attention mechanism is used to quantify the closeness between users and their friends. We considers both individual similarity and their close friends similarity by jointly optimize them in a single objective function. Quantification of close friends improves the generalization ability of the FCUM. Due to the expensive costs of labeling new match users for training FCUM, we also design a bi-directional matching strategy. Experiments on real datasets illustrate that FCUM outperforms other state-of-the-art methods that only consider the individual similarity.

Community detection is a complex and meaningful process, which plays an important role in studying the characteristics of complex networks. In recent years, the discovery and analysis of community structures in complex networks has attracted the attention of many scholars, and many community discovery algorithms have been proposed. Many existing algorithms are only suitable for small-scale data, not for large-scale data, so it is necessary to establish a stable and efficient label propagation algorithm to deal with massive data and complex social networks. In this paper, we propose a novel label propagation algorithm, called WRWPLPA (Parallel Label Propagation Algorithm based on Weight and Random Walk). WRWPLPA proposes a new similarity calculation method combining weights and random walks. It uses weights and similarities to update labels in the process of label propagation, improving the accuracy and stability of community detection. First, weight is calculated by combining the neighborhood index and the position index, and the weight is used to distinguish the importance of the nodes in the network. Then, use random walk strategy to describe the similarity between nodes, and the label of nodes are updated by combining the weight and similarity. Finally, parallel propagation is comprehensively proposed to utilize label probability efficiently. Experiment results on artificial network datasets and real network datasets show that our algorithm has improved accuracy and stability compared with other label propagation algorithms.

Object detection is an essential technology in the computer vision domain and plays a vital role in intelligent transportation. Intelligent vehicles utilize object detection on images for environment perception. This work develops a target detection algorithm based on deep learning technologies, particularly convolutional neural networks and neural network modeling. Building on the analysis of the traditional Haar-like vehicle recognition algorithm, a vehicle recognition algorithm based on a convolutional neural network with fused edge features (FE-CNN) is proposed. The experimental results demonstrate that FE-CNN improves the recognition precision and the model’s convergence speed through a simple and effective edge feature fusion method. In the experiment conducted using real traffic scene for vehicle recognition, the developed algorithm achieves a 99.82% recognition rate in efficient time, demonstrating the capability for real-time performance and accurate target detection.

Many real-world problems can be abstracted into graph classification problems. Recently, graph convolutional networks have achieved great success in the task of node classification and link prediction. However, when using graph convolution network to process the task of graph classification, either global topology information or local information is ignored. Therefore, designing graph convolutional networks to improve the accuracy of graph classification has attracted more and more attention. Inspired by the use of convolutional neural networks to process graph-structured data, we put forward a new spatial convolutional neural network architecture for graph classification. To be specific, we first design a comprehensive weighting method to measure the significance of vertices in the graph based on multiple indicators to choose the central node sequence. Then, the normalization process of the graph is realized by constructing the same size neighborhood graphs for the central vertices. After that, the structural characteristics of the graph are extracted from both local and global aspects. Finally, the tensors obtained after the above steps are respectively input into the following two spatial convolutional neural network architectures to perform classification, one is a simple CNN structure, which has only two convolution layers, one dense layer and one softmax layer. The other is to modify the architecture of CNN, and the channel concatenation layer is introduced to determine the classification result of the entire graph according to the category of the neighborhood graphs. Experimental results on two kinds of real-world datasets, bioinformatics and social network datasets, indicate that our approach obtains competitive results and is superior to some classic kernels and similar deep learning-based algorithms on 6 out of 8 benchmark data sets.

Crowd management is a considerable challenge in many countries including Saudi Arabia, where millions of pilgrims from all over the world visit Mecca to perform the sacred act of Hajj. This holy ritual requires large crowds to perform the same activities during specific times, which makes crowd management both critical and difficult. Without proper crowd management and control, the occurrence of disasters such as stampedes, suffocation, and congestion becomes highly probable. At present, the internet of things (IoT) and its enabling technologies represent efficient solutions for managing and controlling crowd, minimizing casualties, and integrating different intelligent technologies. Moreover, IoT allows intensive interaction and heterogeneous communication among different devices over the internet, thereby generating big data. This paper proposes an intelligent IoT approach for crowd management with congestion avoidance in the Mina area, which is located in the holy city of Mecca. The approach implements a learning mechanism that classifies pilgrims based on the collected data and exploits the advantages of both IoT and cloud infrastructures to monitor crowds within a congested area, identify evacuation paths for pilgrims and guide the pilgrims to avoid congestion in real time. Moreover, the approach attempts to maximize crowd safety based on realistic scenarios by controlling and adapting pilgrim movements according to the characteristics of the possible hazards, pilgrim behavior, and environmental conditions. We evaluated our proposed approach by performing simulations based on real data sets and scenarios.

Conversation generation is an important natural language processing task and has attracted much attention in recent years. The realization of the conversation model is also of great significance to the field of social computing, helping to build artificial intelligence robots on social networks. The open domain conversation model is fundamentally data-driven, which can be roughly divided into retrieval models and generation models. Although remarkable progress has been achieved in recent years, it is still difficult to get responses that are grammatically and semantically appropriate. We propose the Rerank of Retrieval-based and Transformer-based Conversation model (RRT), a novel conversation model that combines the retrieval model with the generation model for the purpose of obtaining context–appropriate response. The context–response pairs with the highest similarity from training set are retrieved using traditional retrieval method, and further ranked to obtain the retrieval candidate response. We replaced the traditional sequence-to-sequence models for conversation generation by the transformer model and achieved better results with less training time. Finally, the post-reranking module is used to rank the retrieved candidate and the generated one to obtain the final response. We conducted detailed experiments on two datasets and the results show that compared with the traditional generation model, our model has a significant improvement in each metric, and the training time is decreased by a factor of 5. Furthermore, our model is more informative and relevant to the input context than the retrieval model.

Mass gatherings (such as Hajj/Umrah), owing to their immensity, often present a variety of difficulties to the attendees. To have a comprehensive understanding of the difficulties as well as their potential remedies, surveying a good number of attendees is unavoidable and this can be facilitated through engaging data collectors. A crucial part here is identifying the integrity of the data collectors, which is yet to be explored in the literature to the best of our knowledge. To address this gap, in this study, we first perform a mass-scale data collection over Hajj/Umrah pilgrims through online (n = 236) and in-person (n = 752) surveys, where we cover a substantial part (n = 712) through paid data collectors (n = 53). We critically investigate data collection activities of the data collectors through focused group discussions involving expert reviewers. We explore two computing techniques to unveil the integrity of the data collectors from two different perspectives. Our study finds out influential (religious and socio-geographical) aspects that impact the data collection process. Besides, we find that the collaborative participation of expert reviewers is obligatory to scrutinize the data collectors’ integrity. Additionally, our explored computing techniques, namely conflict analysis and learning-based analysis, can identify up to 74% and 99% of the unreliable data collectors respectively. We observe that, although these computing-based filtering can indicate the integrity up to a certain level, human-in-the-loop is unavoidable for concluding on the integrity. To the best of our knowledge, this is the first study of its kind (i.e., integrity analysis of the data collectors) in the literature.

With the fast development of various computing paradigms, the amount of data is rapidly increasing that brings the huge storage overhead. However, the existing data deduplication techniques do not ...

In this article, first a comprehensive study of the impact of term weighting schemes on the topic modeling performance (i.e., LDA and DMM) on Arabic long and short texts is presented. We investigate six term weighting methods including Word count method (standard topic models), TFIDF, PMI, BDC, CLPB, and CEW. Moreover, we propose a novel combination term weighting scheme, namely, CmTLB. We utilize the mTFIDF that takes into account the missing terms and the number of the documents in which the term appears when calculating the term weight. For further robust term weight, we combine mTFIDF with two weighting methods. We evaluate CmTLB against the studied weighting schemes by the quality of the learned topics (topic visualization and topic coherence), classification, and clustering tasks. We applied weighting schemes to Latent Dirichlet allocation (LDA) and Dirichlet multinomial mixture (DMM) on eight Arabic long and short document datasets, respectively. The experiment results outline that appropriate weighting schemes can effectively improve topic modeling performance on Arabic texts. More importantly, our proposed CmTLB significantly outperforms the other weighting schemes. Secondly, we investigate whether the Arabic stemming process can improve topic modeling performance. We study the three approaches of Arabic stemming including root-based, stem-based, and statistical approaches. We also train topic models with weighting schemes on documents after applying four stemmers related to different stemming approaches. The results outline that applying the stemming process not only reduces the dimensionality of term-document matrix leading to fast estimation process, but also show enhancement of topic modeling performance both on short and long Arabic documents. Moreover, Farasa stemmer achieves the highest performance in most cases, since it prevents the ambiguity that may happen because of the blind removal of the affixes such as in root-based or stem-based stemmers.

The performance of differential evolution (DE) algorithm highly depends on the selection of mutation strategy. However, there are six commonly used mutation strategies in DE. Therefore, it is a challenging task to choose an appropriate mutation strategy for a specific optimization problem. For a better tackle this problem, in this paper, a novel DE algorithm based on local fitness landscape called LFLDE is proposed, in which the local fitness landscape information of the problem is investigated to guide the selection of the mutation strategy for each given problem at each generation. In addition, a novel control parameter adaptive mechanism is used to improve the proposed algorithm. In the experiments, a total of 29 test functions originated from CEC2017 single-objective test function suite which are utilized to evaluate the performance of the proposed algorithm. The Wilcoxon rank-sum test and Friedman rank test results reveal that the performance of the proposed algorithm is better than the other five representative DE algorithms.

Cloud computing is widely used for its powerful and accessible computing and storage capacity. However, with the development trend of Internet of Things (IoTs), the distance between cloud and terminal devices can no longer meet the new requirements of low latency and real-time interaction of IoTs. Fog has been proposed as a complement to the cloud which moves servers to the edge of the network, making it possible to process service requests of terminal devices locally. Despite the fact that fog computing solves many obstacles for the development of IoT, there are still many problems to be solved for its immature technology. In this paper, the concepts and characteristics of cloud and fog computing are introduced, followed by the comparison and collaboration between them. We summarize main challenges IoT faces in new application requirements (e.g., low latency, network bandwidth constraints, resource constraints of devices, stability of service, and security) and analyze fog-based solutions. The remaining challenges and research directions of fog after integrating into IoT system are discussed. In addition, the key role that fog computing based on 5G may play in the field of intelligent driving and tactile robots is prospected.

The Industrial Internet of Things (IIoT) plays an important role in the development of smart factories. However, the existing IIoT systems are prone to suffering from single points of failure and unable to provide stable service. Meanwhile, with the increase of node scale and network quantity, the maintenance cost presents to be higher. Such a disadvantage can be effectively compensated by the features such as security, privacy, non-tamperability and distributed deployment supported by the blockchain. In this paper, first, an intelligent manufacturing security model based on blockchain was proposed. Due to the high power consumption and low throughput of the traditional blockchain, IoT devices with limited power consumption can not work independently. Therefore, in this paper, a new Merkle Patricia tree (MPT) was adopted to extend the blockchain structure and provide fast query of node status. Second, since the MPT does not support concurrent operation and the data operation performance deteriorates with high data volume, a lock-free concurrent and cache-based Merkle Patricia tree was proposed (CMPT) to support lock-free concurrent data operation, which can improve the data operation efficiency in multi-core system. The experimental results indicate that, compared with the original MPT, the CMPT proposed in this paper effectively reduced the time complexity of data insertion and data query and improved the speed of block construction and data query.

Handwriting recognition remains a challenge in the machine vision field, especially in optical character recognition (OCR). The OCR has various applications such as the detection of handwritten Farsi digits and the diagnosis of biomedical science. In expanding and improving quality of the subject, this research focus on the recognition of Farsi Handwriting Digits and illustration applications in biomedical science. The detection of handwritten Farsi digits is being widely used in most contexts involving the collection of generic digital numerical information, such as reading checks or digits of postcodes. Selecting an appropriate classifier has become an issue highlighted in the recognition of handwritten digits. The paper aims at identifying handwritten Farsi digits written with different handwritten styles. Digits are classified using several traditional methods, including K-nearest neighbor, artificial neural network (ANN), and support vector machine (SVM) classifiers. New features of digits, namely, geometric and correlation-based features, have demonstrated to achieve better recognition performance. A noble class of methods, known as deep neural networks (DNNs), is also used to identify handwritten digits through machine vision. Here, two types of introduce its expansion form, a convolutional neural network (CNN) and an auto-encoder, are implemented. Moreover, by using a new combination of CNN layers one can obtain improved results in classifying Farsi digits. The performances of the DNN-based and traditional classifiers are compared to investigate the improvements in accuracy and calculation time. The SVM shows the best results among the traditional classifiers, whereas the CNN achieves the best results among the investigated techniques. The ANN offers better execution time than the SVM, but its accuracy is lower. The best accuracy among the traditional classifiers based on all investigated features is 99.3% accuracy obtained by the SVM, and the CNN achieves the best overall accuracy of 99.45%.

With XML becoming a promising standard for data storage, describing, transferring and exchanging information on the Internet, data security and privacy protection of XML become the focus of research in recent years. In order to achieve the authorization of legitimate user and ensure the secure access to sensitive information, in this paper, in the context of cloud storage, with the purpose of sharing sensitive XML information, a polynomial authorization scheme with Kerberos authentication was proposed, which was based on the users' access purpose and privacy policy. In this scheme, first, Kerberos authentication was used to identify the user, and then the polynomial whose coefficients were from the leaf node address was used to complete the authorization of user. For the legitimate user, under the interaction of authorization polynomials and the global structure view, authorization matrix is generated dynamically, its temporary and dynamic characteristics greatly improves the security of the system. Finally, with the help of authorization matrix and auxiliary information tables, security queries were successfully completed. The experimental results show that the scheme not only effectively protects XML sensitive data, but also reduces the server's storage pressure, at the same time it is beneficial to the rapid search and information positioning.

Location-based Service has become the fastest growing activity related service that people use in their daily life due to the boom of location-aware mobile devices. In edge computing along with the benefits brought by LBS, privacy preservation becomes a more challenging issue because of the nature of the paradigm, in which peers may cooperate with each other to collect and analyze user's location data. To avoid potential information leakage and usage, user's exact location should not be exposed to the edge node. In this paper, we propose a stochastic location privacy protection scheme for edge computing, in which the geographical distribution of surrounding users is obtained by analyzing proposed long-term density map and short-term density map. The cloaking scheme transfers user's exact location to a cloaked location to satisfy predefined probability of having k-users in that area. Our scheme does not reveal any exact location information, thus it is practicable for the real scenario when edge computing is honest but curious. Extensive experimental results are conducted to verify the efficiency and effectiveness of our method. By varying the privacy protection requirements, the corresponding performance have been examined and discussed.

Most current automatic summarization methods are for English texts. The distinction between words in Chinese text is large, the types of parts of speech are many and complex, and polysemy or ambiguous words appear frequently. Therefore, compared with English text, Chinese text is more difficult to extract useful feature words. Due to the complex syntax of Chinese, there are currently relatively few automatic summarization methods for Chinese text. In the past, only the important sentences in the original text can be selected and simply arranged to obtain a summary with chaotic sentences and insufficient coherence. Meanwhile, because Chinese short text usually contains more redundant information and the sentence structure is not neat, we propose a topic-based automatic summary method for Chinese short text. Firstly, a key sentence selection method is proposed combining topic words and TF-IDF to obtain the score of each text corresponding to the topic in the original text data. Then the sentence with the highest score as the topic sentence of the topic is selected. Considering that the short text of Weibo may contain a lot of irrelevant information and sometimes even lack some important components of topic, three retouching mechanisms are proposed to improve the conciseness, richness and readability of topic sentence extraction results. We validate our approach on natural disaster and social hot event datasets from Sina Weibo. The experimental results show that the polished topic summary not only reflects the exact relationship between topic sentences and natural disasters or social hot events, but also has rich semantic information. More importantly, we can almost grasp the basic elements of natural disaster or social hot event from the topic sentence, so as to help the government guide disaster relief or meet the needs of users for quickly obtaining information of social hot events.

One of the most challenging problems during disasters involving crowds is pedestrian evacuation. It is very important in such situations to improve survival rates by getting all or most of the people out in the shortest possible time. Technological intervention through augmenting the evacuation process using an unmanned aerial vehicle (UAV) exhibits great potential in improving survival rates, but the exploration of this method is still in its infancy. Therefore, this study explores the potential of utilizing UAVs for crowd management during emergency evacuations. We conducted a rigorous study, using a simulation model featuring four UAVs and differing numbers of pedestrians, with use of two UAV guidance approaches: partial guidance and full guidance. The experimental results suggest that exploiting UAVs in crowd evacuation and following the partial guidance approach can lead to a more efficient evacuation process.

Today's Internet of Things (IoT) and its enabling technologies offer efficient solutions for utilizing available resources and integrating different techniques. In high-risk environments, IoT has already shown to have significant potential to provide safe, reliable, and efficient solutions. This paper proposes a hybrid emergency evacuation approach using IoT and combines it with cloud computing. The proposed approach attempts to utilize the advantages of both infrastructures (IoT and cloud) to calculate evacuation paths in real time during emergency evacuation. It attempts to maximize the safety of the suggested evacuation paths by adapting to the characteristics of the hazard, evacuees’ behavior, and environmental conditions. Our approach covers two emergency navigators: A Localized Emergency Navigator (LEN) and a High-Risk Emergency Navigator (HREN). Depending on some defined emergency factors, our approach decides which navigator to execute. It also handles diversified evacuation issues, such as when people get locked in a safe, however, a dead-end area of a building under emergency. Performance evaluation shows that the proposed approach achieves better outcomes in terms of survival rate and evacuation efficiency compared to the traditional algorithm.