Search

As a class of randomized learner model, stochastic configuration networks (SCNs) have been successfully applied in a few data analytics tasks. Given the industrial big data modeling tasks, however, the original SCNs potentially lead to excessive training time. To this end, this paper extends SCNs to a hybrid parallel version, termed HPSCNs. In the hybrid parallel learning algorithm, two SCNs are synchronously constructed. The difference between the two of them is that one employs a point-incremental algorithm, and another one adopts a block-incremental algorithm. Moreover, a data parallel method is established based on a dynamic block strategy to accelerate the establishment of candidate node pool for each SCN. Additionally, this paper proposes an adaptive hyperparameter adjustment method, which allows the hyperparameters in the supervisory mechanism to be automatically adjusted along with the learning process. Comparative experiments are first conducted through four large-scale benchmark datasets, followed by the fully discussion on the algorithm parameters. Finally, a practical industrial application case is made, where a grinding particle size soft sensor is developed based on HPSCN, showing the effectiveness of the proposed algorithm.

Cloud contamination is a common phenomenon in the optical remote sensing field, which limits their application in land surface studies and causes the waste of satellite images. This letter presents a new framework for removing thin clouds from visible images based on multiscale dark channel prior (MDCP). The cloud removal of cloudy images (target images) is carried out with the assistance of a different temporal cloudless image (reference image) from the perspective view of multiscale transform (MST). In order to make it more suitable for the application of thin cloud removal, two improvements are made to this traditional fusion method. For one thing, a dark channel prior module is integrated into the low-frequency component of the target image in the framework of MST. For another, we choose the weighted average for high-frequency components and sparse representation (SR) for low-frequency components as fusion rules. After the fusion process, the modified Laplacian sharpening whose model is optimized is carried out. The performance of MDCP was evaluated with both simulated and real cloudy images. Experimental results show that the proposed MDCP has a good performance.

Conflict is an inevitable social phenomenon and the analysis of it can effectively resolve disputes, improve the position of decision makers (DMs), forecast compromise solutions as well. At present, the Graph Model for Conflict Resolution (GMCR), a completely non-quantitative decision support system (DSS) based on DM’s ordinal preference information, is constructed to combat conflict analysis complicated by multiple participant or multiple criteria or both. The purpose of the study is to introduce an overview of GMCR in conflict analyzing through bibliometrics. In order to achieve this goal, a systematic review of articles in leading journals of Web of Science Core Collection (WoSCC) during 1987–2019 is posed, referring to the distribution of countries, institutions, authors, subjects and journals, research topics and hotspots exploration as well as frontiers prediction, by utilizing VOSviewer and CiteSpace. The contributions of this study are not only providing a handy method to grasp generalized scientific research situation, but also demonstrating status quo and emerging trends of GMCR for researchers and everyone who interested in.

Radio-frequency fingerprinting (RFF) exploiting hardware characteristics has been employed for device recognition to enhance the overall security. However, the performance unreliability in long-term experiments, channel fading interference, and unauthorized devices verification are three open problems that restrict the development of RFF recognition. To address these issues, a robust RFF extraction scheme based on three corresponding algorithms is studied. For the first problem, a long-term stacking of repetitive symbols (LSRSs) algorithm is proposed to reduce the acquired signal variance, which contributes to the identification accuracy and long-term stability. For the second issue, we propose an artificial noise adding (ANA) algorithm to enhance the recognition robustness through regularization and channel adaptation. For the third issue, a verification algorithm based on the generative Gaussian probabilistic linear discriminant analysis (GPLDA) model is developed to handle unauthorized devices. Our robust RFF extraction scheme is verified in the experiments with 54 CC2530 ZigBee devices. It enables reliable node identification with the accuracy of 99.50% in the short rang line-of-sight (SLOS) scenarios for signals collected over 18 months, and 95.52% in the extensive multipath fading experiments. The equal error rate (EER) of the verification experiments with six authorized devices versus six unseen unauthorized devices is as low as 0.63%.

The disadvantage of low resolution and low signal-to-noise ratio (SNR) of thermal infrared (TIR) data makes it difficult to be applied in practice, although it has been studied for decades. In contrast, visible and near infrared (VNIR) and shortwave infrared (SWIR) data have higher resolution and SNR. In this letter, an ensemble extreme learning machine (EELM) is proposed, which is based on the extreme learning machine (ELM), to fuse together VNIR, SWIR, and TIR to improve TIR resolution and SNR. The associated ELM nets are trained with correlated training data which are selected from the random forest method. The temperature estimation cube can be obtained after putting SWIR data and VNIR data into these networks. Finally, the cube structure is adjusted by an ensemble rule to obtain the final temperature estimation results. We test the EELM algorithm on the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) L1B data set and compare it with other scholars’ algorithms. The experimental results show that the EELM algorithm can not only improve the resolution of TIR data, but also estimate the subpixel temperature of TIR more accurately.

Artificial bee colony (ABC) algorithm is a relatively new paradigm of swarm intelligence based optimization technique, which has attracted a lot of attention for its simple structure and good performance. For some complex optimization problems, however, the performance of ABC is challenged due to its solution search equation that has strong explorative ability but poor exploitative ability. To solve this defect, in this work, we propose an improved ABC algorithm by using multi-elite guidance, which has the benefits of utilizing valuable information from elite individuals to guide search while without losing population diversity. First, we construct an elite group by selecting some elite individuals, and then introduce two improved solution search equations into the employed bee phase and onlooker bee phase based on the elite group, respectively. Last, we develop a modified neighborhood search operator by utilizing the elite group as well, which aims to achieve a better tradeoff between explorative and exploitative abilities. To verify our approach, 50 well-known test functions and one real-world optimization problem are used in the experiments, including 22 scalable basic test functions and 28 complex CEC2013 test functions. Seven different well-established ABC variants are involved in the comparison and the results show that our approach can achieve better or at least comparable performance on most of the test functions.

The retrieval of land surface temperature (LST) using thermal infrared (TIR) data is important in many applications. However, TIR data usually suffer from low spatial resolution. We introduce a novel subpixel LST estimation model using the information-transfer-based adaptive ensemble extreme learning machine (IT-AEELM). The proposed method constructs a reliable relationship between subpixel LST and the input high-resolution visible and near-infrared (VNIR) data, short-wave infrared (SWIR) data, and low-resolution TIR data. Based on a detailed analysis of different ground objects, we divide the input data into multiple subsets. Instead of using consistent land surface parameters (LSPs), we utilize different LSPs to characterize the land surface properties in each subset. The VNIR-SWIR-LSPs data and the low-resolution LST are used to train a novel IT-AEELM network, where a feedback ensemble learning scheme is introduced to effectively remove inaccurate estimates. The main difference of the model against existing methods is that it builds a robust architecture at different spatial scales, which provides benefits including lower demand for training data, more rapid and accurate acquisition of subpixel LST, and better adaption to heterogeneous land surface. Numerical experiments demonstrate that the proposed method significantly improves the accuracy of subpixel LST compared with the state-of-the-art algorithms.

Artificial bee colony (ABC) algorithm is one of the branches of swarm intelligence. Several studies proved that the original ABC has powerful exploration and weak exploitation capabilities. Therefore, balancing exploration and exploitation is critical for ABC. Incorporating knowledge in intelligent optimization algorithms is important to enhance the optimization capability. In view of this, a novel ABC based on knowledge fusion (KFABC) is proposed. In KFABC, three kinds of knowledge are chosen. For each kind of knowledge, the corresponding utilization method is designed. By sensing the search status, a learning mechanism is proposed to adaptively select appropriate knowledge. Thirty-two benchmark problems are used to validate the optimization capability of KFABC. Results show that KFABC outperforms nine ABC and three differential evolution algorithms.

Artificial bee colony (ABC) and its most modifications use a probability method to select good food sources (called solutions) in the onlooker bee search phase. However, the probability selection does not work with increasing of iterations, because the fitness values cannot be used to distinguish two different solutions. In order to tackle this problem, this paper proposes a new ABC (called NSABC), in which a new selection method based on neighborhood radius is used. Unlike the probability selection in the original ABC, NSABC chooses the best solution in the neighborhood radius to generate offspring. Based on the neighborhood radius, two new solution search strategies are modified. The scout bee search phase is improved by using opposition-based learning and the neighborhood radius. To evaluate the search ability of NSABC, there are 22 benchmark problems used in the experiments. Performance comparison shows NSABC achieves better results than five other ABC algorithms.

Most image fusion algorithms based on hyperspectral unmixing perform worse with the lower spatial resolution of hyperspectral image (HSI) for the reason that the estimated endmembers and abundance deviate from the truth value. Therefore, it is more meaningful to unmix the low spatial resolution hyperspectral image (LRHSI) accurately, which is also helpful to improve the image fusion performance. In order to enhance the spatial resolution of LRHSI, this article proposes an alternating direction iterative nonnegative matrix factorization (ADINMF) based on linear hyperspectral unmixing algorithm. It takes multispectral image as a constraint to improve the spatial resolution of LRHSI. First, we use blind source separation to initialize the endmember and abundance of hyperspectral and multispectral images, respectively. Then, we alternately update the endmembers and abundance in the framework of nonnegative matrix factorization by multiplication iterative algorithm. The updated endmembers and abundance are constrained to each other. We compare the experimental results of simulated dataset and three groups of real datasets. Experimental results show that the proposed method not only accurately extracts the endmembers of LRHSI, but also obtains a significant fusion performance improvement.

Jumping rope is a very efficient aerobic exercise, during which the exerciser can maintain an efficient heart rate range (heart rate value to 140 ∼ 160bpm). However, at present, people are unable to recognize whether they are in this ideal aerobic state in a timely and accurate manner, and there is no convenient way of showing exercisers how to perform efficient jumping rope exercises. We developed an early prototype, Rope X, an exercise device that communicates the exerciser's heart rate status through a dynamic interactive interface and gives the exerciser a certain degree of guidance, as to the correct use of the jumping rope to improve the efficiency of the jumping rope exercise (aerobic exercise).

With the development of the vehicular network, BusNet based service is of great interest to both bus companies and public users. But in this mode, due to the dynamic characteristics of the vehicle, how to effectively communicate has become a problem. In addition, considering the problem of jumping out of the local optimum, a new route optimization algorithm CSPSO (chaotic serial particle swarm optimization), is proposed. The comprehensive simulation results demonstrate that the proposed method can not only adapt to the dynamic changes of network topology, but also greatly reduce the possibility of falling into local optimum, and improve the communication performance significantly.

Remote sensing images contaminated by clouds cannot be used for target recognition, image classification, and other applications, which leads to a lot of remote sensing data being wasted. We propose a low-rank and sparse constrained dark channel prior for cloud removal in remote sensing image sequence (LRSC-DCP). The sparse and low-rank constraints are used to find the position of thick clouds and remove thick clouds in cloud-contaminated images respectively. The remaining thin clouds can be eliminated by the dark channel prior. We compare the algorithm proposed in this article with the dark channel prior. Experimental results prove that the proposed method has better performance.

Hyperspectral image (HSI) classification methods based on deep learning (DL) algorithms have achieved significant improvements on abundant samples. However, due to the limitation of practically available samples, the difficulty of representative feature extraction from small-sized samples and the loss of subtle diagnostic features in DL iteration results in the accuracy reduction of interclass and intraclass in refined classification, respectively. To address these issues, a water retrieval embedded deep network is proposed in this paper. The relative water content retrieval (RWCR) of the proposed network is embedded as a subnet, which is responsible for extracting subtle diagnostic features of relative water content (RWC) to enhance the representation of features in classification. The experimental results verify the effectiveness of RWCR for improving the interclass and intraclass accuracy in refined classification. Moreover, the superiority of the proposed network is also demonstrated in comparison with state-of-the-art methods.

The cloud-contaminated phenomenon in the field of remote sensing has a serious impact on image processing so that a large number of images are unusable. To achieve cloud removal for single visible image, we propose a novel method based on modified dark channel prior with multiple scale (MDCPMS). In the structure of multiple scale, the cloudy image is firstly decomposed into high-frequency and low-frequency components. The former is uniformly amplified to enhance its weak contour, and the latter is processed by modified dark channel prior (DCP), whose estimation of atmospheric light is optimized for better cloud removal. Finally, the cloud-removed image is obtained through multi-scale reconstruction. Experimental results show that the proposed MDCPMS obtains a significant performance with slightest color distortion and is closest to the corresponding real image, compared with DCP and nonlocal method.

We address the problem of jointly denoising and super resolving neuromorphic events, a novel visual signal that represents thresholded temporal gradients in a space-time window. The challenge for event signal processing is that they are asynchronously generated, and do not carry absolute intensity but only binary signs informing temporal variations. To study event signal formation and degradation, we implement a display-camera system which enables multi-resolution event recording. We further propose Event- Zoom, a deep neural framework with a backbone architecture of 3D U-Net. EventZoom is trained in a noise-to-noise fashion where the two ends of the network are unfiltered noisy events, enforcing noise-free event restoration. For resolution enhancement, EventZoom incorporates an event-to- image module supervised by high resolution images. Our results showed that EventZoom achieves at least 40 × temporal efficiency compared to state-of-the-art (SOTA) event denoisers. Additionally, we demonstrate that EventZoom enables performance improvements on applications including event-based visual object tracking and image reconstruction. EventZoom achieves SOTA super resolution image reconstruction results while being 10× faster.

This study aims to explore the current state of AI education in massive open online courses (MOOCs). We screened 128 sample MOOC courses from 10 learning platforms and coded information on course overviews, content outlines, recommended materials and teaching methods as a way to understand the content and teaching methods of AI education on MOOC platforms. The results show that: (1) Coursera and Edx platforms offer the largest number of ESD-related courses, the organizations offering MOOCs are mostly companies and universities, and the number of AI education-related courses started from 2012 to the present has increased year by year. The average number of participants is exceeds 120,000, and AI courses are currently very popular; (2) Core algorithms and topics related to typical applications are the main content of AI education; (3) Online procedural teaching, task-driven teaching and online collaborative learning are the most used pedagogical methods in MOOCs at this stage. However, pedagogical approaches such as scenario-based learning and project-based learning are not yet fully utilized in the MOOC environment. Our findings suggest that MOOC AI education in research a practical and innovative approach in preparing and learning AI education-related MOOC courses.

Artificial bee colony (ABC) algorithm is a popular optimization technique with strong search ability. Although ABC has the ability to handle complex optimization problems, it suffers from weak exploitation and slow convergence. In order to tackle this issue, a new ABC variant based on multiple search strategies and dimension selection (ABC-MSDS) is proposed in this paper. Firstly, multiple search strategies based on dual strategy pool are designed. Compared to other existing ABC with multiple search strategies, our approach constructs two strategy pools for employed and onlooker bees, respectively. Secondly, a new dimension selection method is used to replace the random dimension selection in the standard ABC. In the search process, each dimension is chosen one by one in terms of the quality of offspring. Finally, a modified scout bee phase is employed to accelerate the search. Experimental study is conducted on classical benchmark problems and CEC 2013 shifted and rotated problems. The performance of ABC-MSDS is compared with several recently published ABC variants. Computational results demonstrate the effectiveness of our approach.

Matrix decomposition is an important technology in collaborative filtering recommendation algorithm. Based on SVD, some improvements, e.g., RSVD algorithm with offset term and regular term, and SVD++ algorithm with explicit and implicit feedback, was proposed. Considering the influence of time factor on the recommendation process, we discussed a time factor to the algorithms, which were associated with the users, the projects and the interaction offset. Based on the different time decay characteristics of video and users, an algorithm with exponential correction called timeSVD++EXP is proposed. The results show that the proposed algorithm can improve the prediction accuracy of the model when root mean square error is used as the evaluation index. And the prediction accuracy of the model is further increased by considering the time factor and exponential correction.

With the rapid development of computer technology, higher requirements are put forward for the design of game software. As a typical example, artificial intelligence technology is one of the main technologies in the development and application of game software in recent years, which can meet the requirements of realtime characteristics and bring more rich experience for game players. This paper mainly discusses the application of AI technology from game engine, character movement and path planning.

In this paper, we devise a framework to infer a smartphone user’s location and walking routes in an indoor environment, using only the information from inertial measurement unit (IMU) sensors like gyroscope and accelerometer. To overcome the shortcoming of estimation drift over time in common PDR (pedestrian dead reckoning)-IMU systems, we propose a map-aided system which uses the map elements to help correct the user’s position. We generate a map based on the environment parameters and a map matching algorithm is applied to find the most likely location of the user. The reading from the IMU sensors contains amounts of noise when user is walking, therefore we propose an edge detection algorithm based on the pruned exact linear time (PELT) model to smooth the piece-wise signals and identify the time frame when the user is making a turn. We evaluate our system when the smartphone is held either in the user’s hand or in the backpack, and the system is able to give the correct walking path in both cases.

Global system for mobile communications (GSM) is one of the most widely used communication standards in the world today, which still has a large number of users, so it is of great security significance to identify devices operating in a GSM network. This paper proposes a novel radio frequency fingerprinting (RFF) based device identifications method for mobile phones. A differential constellation trace figure (DCTF) physical layer RFF extraction and convolutional neural network (CNN) based classification scheme is designed to identify accessing mobile phones. Theoretical analysis shows that differential process of GSM signal can effectively reflect the characteristics of RFF from different phones. Compared with the existing RFF identification methods, CNN based classification can identify the DCTF of different devices with low complexity and high accuracy. Furthermore, the proposed DCTF-CNN method is robust to different device locations and GSM parameters. Experimental results show that the accuracy of the proposed DCTF-CNN method can reach 92.97% and 99.77% with SNR at 25 dB and 50 dB for 6 mobile phones.

With the rapid development of information technology and the increasing popularity of information and communication technology, the information age has come. Enterprises must adapt to changes in the times, introduce network and computer technologies in a timely manner, and establish more efficient and reasonable information systems and platforms. Large-scale information system construction is inseparable from related audit work, and network security risks have become an important part of information system audit concerns. This paper analyzes the objectives and contents of information system audits under the background of network information security through theoretical analysis, and on this basis, proposes how the IS audit work will be carried out.

The pandemic of COVID-19 has caused millions of infections, which has led to a great loss all over the world, socially and economically. Due to the false-negative rate and the time-consuming of the conventional Reverse Transcription Polymerase Chain Reaction (RT-PCR) tests, diagnosing based on X-ray images and Computed Tomography (CT) images has been widely adopted. Therefore, researchers of the computer vision area have developed many automatic diagnosing models based on machine learning or deep learning to assist the radiologists and improve the diagnosing accuracy. In this paper, we present a review of these recently emerging automatic diagnosing models. 70 models proposed from February 14, 2020, to July 21, 2020, are involved. We analyzed the models from the perspective of preprocessing, feature extraction, classification, and evaluation. Based on the limitation of existing models, we pointed out that domain adaption in transfer learning and interpretability promotion would be the possible future directions., In ICRAI 2020: 2020 6th International Conference on Robotics and Artificial Intelligence

In this paper, we propose a novel method called Residual Steps Network (RSN). RSN aggregates features with the same spatial size (Intra-level features) efficiently to obtain delicate local representations, which retain rich low-level spatial information and result in precise keypoint localization. Additionally, we observe the output features contribute differently to final performance. To tackle this problem, we propose an efficient attention mechanism - Pose Refine Machine (PRM) to make a trade-off between local and global representations in output features and further refine the keypoint locations. Our approach won the 1st place of COCO Keypoint Challenge 2019 and achieves state-of-the-art results on both COCO and MPII benchmarks, without using extra training data and pretrained model. Our single model achieves 78.6 on COCO test-dev, 93.0 on MPII test dataset. Ensembled models achieve 79.2 on COCO test-dev, 77.1 on COCO test-challenge dataset. The source code is publicly available for further research at https://github.com/caiyuanhao1998/RSN/.

Most graph-network-based meta-learning approaches model instance-level relation of examples. We extend this idea further to explicitly model the distribution-level relation of one example to all other examples in a 1-vs-N manner. We propose a novel approach named distribution propagation graph network (DPGN) for few-shot learning. It conveys both the distribution-level relations and instance-level relations in each few-shot learning task. To combine the distribution-level relations and instance-level relations for all examples, we construct a dual complete graph network which consists of a point graph and a distribution graph with each node standing for an example. Equipped with dual graph architecture, DPGN propagates label information from labeled examples to unlabeled examples within several update generations. In extensive experiments on few-shot learning benchmarks, DPGN outperforms state-of-the-art results by a large margin in 5% $\sim$ 12% under supervised setting and 7% $\sim$ 13% under semi-supervised setting. Code will be released., Comment: Accepted by CVPR 2020

Temporal relation classification, an important branch of relation extraction, aims to identify the time sequence among events. Currently, Shortest Dependency Path (SDP) is widely used in various kinds of neural network models to capture the crucial information from sentences. However, while eliminating irrelevant words in event sentences, SDP will miss some useful information, e.g., time expressions. To address the above issue, we propose a neural network method incorporating the temporal cues to AC-GCN (Augmented Contextualized Graph Convolutional Network) to classify temporal relations. Firstly, we introduce the semantic role labeling and heuristic rules to extract the time expressions corresponding to event triggers and other words in SDPs, respectively. Then, the SDP with time expression (i.e., T-SDP) is encoded by a Bi-LSTM with the parameter sharing mechanism and fed into GCN to classify temporal relations. The experimental results on TimeBank-Dense show that our proposed model outperforms all baselines significantly.

As a relatively new paradigm of bio-inspired optimization techniques, artificial bee colony (ABC) algorithm has attracted much attention for its simplicity and effectiveness. However, the performance of ABC is not satisfactory when solving some complex optimization problems. To improve its performance, we propose a novel ABC variant by designing a dynamic multi-population scheme (DMPS). In DMPS, the population is divided into several subpopulations, and the size of subpopulation is adjusted dynamically by checking the quality of the global best solution. Moreover, we design two novel solution search equations to maximize the effectiveness of DMPS, in which the local best solution of each subpopulation and the global best solution of the whole population are utilized simultaneously. In the experiments, 32 widely used benchmark functions are used, and four well-established ABC variants are involved in the comparison. The comparative results show that our approach performs better on the majority of benchmark functions.

As one of the most thrilling tasks of natural language understanding (NLU), intent classification in a dialogue system has received a great deal of attention in both industry and academia. The major limiting factor on intent classification is the lack of tagged data. To solve it, in this paper, we propose a conditional sequence generative adversarial network (cSeq-GAN) for intent classification of short-spoken language, in which we simultaneously train a generative model and a discriminative model for two tasks, one to distinguish the generated text from the real spoken one, while the other to predict its intent category. More reliable tagged data obtained by the generator greatly improves the performance of the intent classification task. Extensive experiments on both Air Travel Information System (ATIS) and our selling robot dialogue system for insurance industries demonstrate that our cSeq-GAN achieves competitive classification accuracy with other state-of-art methods of text classification.

Firefly algorithm (FA) is an effective optimization technique based on swarm intelligence, which has been successfully applied to various practical engineering problems. In this paper, a new dynamic FA (called NDFA) is proposed for demand estimation of water resources in Nanchang city of China. First, a dynamic parameter strategy is used to avoid manually adjusting the step factor. Second, three estimation models in different forms (linear, exponential and hybrid) are developed in terms of the historical water use and local economic structure. Third, normalization method is utilized to eliminate the influences of different units of data. In the experiments, water use in Nanchang city from 2003 to 2015 is considered as a case study. The data from 2003 to 2012 are used for finding the optimal weights of the models, and the rest of data (2013–2015) are applied to test the models. Computational results show that all five FA variants can achieve promising solutions. The proposed NDFA obtains better performance than four other FA variants, and its prediction accuracy is up to 97.91%. Finally, the water demand in Nanchang city from 2017 to 2020 is predicted.

Cloud model and D-S theory have been widely used in uncertainty reasoning. Meanwhile, modal strain energy and Inner Product Vector are also utilized as damage-sensitive features to detect structural damage. In this paper, a new structural damage identification approach is proposed based on Dempster-Shafer theory and cloud model. Cloud models were created to make uncertainty reasoning in damage structures by modal strain energy and the Inner Product Vector of acceleration. Then the results of the two methods were combined by using the Dempster-Shafer theory. Due to the classical D-S theory involves counter – intuitive behavious when the high conflicting evidences exists, the distance function was introduced to correct the conflict factor K and combine the evidences. Moreover, a model of simple beam was created to verify the feasibility and accuracy for the single-damage and the multiple-damage. The effects of noise on damage detection were investigated simultaneously. The results show that the method has strong anti-noise ability and high accuracy.

In recent years, as an effective global optimization technique, artificial bee colony (ABC) algorithm has attracted increasing attention for its good performance yet easy implementation. However, the solution search equation of ABC emphasizes exploration over exploitation, which greatly affects the convergence speed and accuracy of ABC. To solve this problem, many ABC variants have been developed to enhance exploitation by using the superior individuals. However, the concept of neighborhood topology has rarely been considered, which has significant effect on the dissemination of search information among individuals. Hence, in this work, we propose a new ABC variant based on multiple neighborhood topologies (named ABC-MNT). In ABC-MNT, three kinds of neighborhood topologies are used for different individuals to perform diverse abilities of disseminating search information, contributing to a better balance between exploration and exploitation. Considering the characteristics of different neighborhood topologies, three modified solution search equations are assigned to the three neighborhood topologies, respectively, to generate offspring. Furthermore, to preserve search experience of the scout bee phase, the global neighborhood search and opposition-based learning techniques are used. Extensive experiments are carried out on two widely used test suites, and eight well-established ABC variants and five other evolutionary algorithms are included in the performance comparison. The comparison results verify that ABC-MNT can obtain better or at least comparable performance on most of the benchmark functions.

With the large-scale development and construction of urban underground spaces, the safety monitoring of foundation pit construction has gained much attention. Current safety monitoring of foundation pits is often achieved through manual onsite measurements with complex equipment and relies on complicated methods, which are labour intensive, time consuming, and tend to ignore the risk of accidental collapses caused by serious local deformation. To address this issue, this study develops a rapid safety monitoring and analysis method for foundation pit construction using Unmanned Aerial Vehicle (UAV) images. The safe inclination angle of the foundation pit slope was proposed as the safety-monitoring index. Taking the images of the foundation pit captured by the UAV as input, point cloud reconstruction and surface fitting were carried out. The local deformation distribution was introduced to evaluate the local safety state of the foundation pit. The experimental results validate the potential advantages of the method.

Firefly algorithm (FA) is a new optimization technique based on swarm intelligence. It simulates the social behavior of fireflies. The search pattern of FA is determined by the attractions among fireflies, whereby a less bright firefly moves toward a brighter firefly. In FA, each firefly can be attracted by all other brighter fireflies in the population. However, too many attractions may result in oscillations during the search process and high computational time complexity. To overcome these problems, we propose a new FA variant called FA with neighborhood attraction (NaFA). In NaFA, each firefly is attracted by other brighter fireflies selected from a predefined neighborhood rather than those from the entire population. Experiments are conducted using several well-known benchmark functions. The results show that the proposed strategy can efficiently improve the accuracy of solutions and reduce the computational time complexity.

As the sharply development of remote sensing technology, spatial resolution and spectral resolution become much higher in hyperspectral images than before. Commonly, spectral differences often be used in distinguish objects that are difficult to be classified, especially, which share the same color or texture. However, the spectral features are not as unique as we think. In many cases, spectral curves of same materials may be different and, on the contrary, that of different materials may be same. In this condition, false alarm and missing alarm probability are high. To solve this problem, a modified density mixture model is provided. Firstly, each band of data is whitened to remove the correlation between the pixels in order to reduce redundancy. Secondly, the whitened result will be handled by the weighted multivariate normal distribution model. Then, several pixels of each kind of objects are taken to build an spectral library. Finally, Spectral Angle Mapping (SAP) is applied to classification by matching with spectral library. The result demonstrates that objects are classified precisely with low false alarm and missing alarm probability, for the spectral difference of the same kind of objects decreases, and that of different kinds of objects increases compared with the data before processing.

Nowadays, with the development of humanistic consciousness, the residents' travel behavior is becoming more and more important to be considered in urban planning, and has become an important reference for urban traffic construction. The ridesourcing softwares like Uber and Didi have been widely accepted and used. As a model of travel, the ridesourcing has many features, like convenience and flexibility, and the origin and destination of every trip are completely determined by passengers, the running track of vehicles can directly reflect the travel behavior of urban residents. So this paper will study the passengers’ travel behaviour by data mining based on the Didi order data, try to reveal the travel characteristics of urban residents from a macro perspective. This paper focuses on these issues: to analyze the order data from ridesourcing company Didi; to explore the temporal characteristics of passengers’ travel from different angles like the temporal distribution of travel, the distribution of peak hour, time consumption; to analyze the spatial characteristics from the different aspects including the distribution of travel distance, travel hot spots; to propose an optimization model of the taxi stand location by analyzing travel characteristics, which can improve user’s experience of residents by setting up reasonable site layout. Finally we found that the ridesourcing track data can well reveal travel characteristics of urban residents, which could be helpful for the urban planning and road network optimization.

Brain storm optimization (BSO) is a kind of swarm intelligence, which is derived from human brainstorming process. And it has been successfully applied in solving practical problems. As a relatively new algorithm, BSO has the problem of slow convergence speed and being prone to local optimum. In this paper, BSO with role-playing strategy (RPBSO) is proposed. In RPBSO, role-playing strategy is adopted to classify ideas, and idea difference strategy is utilized to generate new ideas. To maintain the diversity of ideas and prevent getting into local optimum, re-initialization operation is adopted as well. RPBSO and several variants of BSO are compared on CEC 2013 benchmark functions. The experimental results show that RPBSO is better than other variants of BSO, which verifies the performance of RPBSO.

Radio frequency (RF) fingerprint has gained wide attention as it takes advantages of inherent characteristics in hardware for identification and verification. However, performance unreliability with long-ago training data and channel fading interference are two open problems that restrict the development of RF fingerprint identification. To address the former issue, we propose a long-term stacking of repetitive symbols algorithm to turn the measurement noise toward standard Gaussian distribution, which contributes to both classification accuracy and long-term stability. For the latter issue, we propose an artificial noise adding algorithm in training stage to enhance the identification robustness. At last, we implement a robust RF fingerprint generation and classification scheme for practical device identification. In the experiment, we used 50 CC2530 ZigBee devices to verify the performance of proposed RF identification scheme. Our scheme achieved an accuracy of 100% with both training and testing data measured in the same day, and enabled reliable node recognition with the accuracy over 99% for testing data collected 18 months later. Further exploration shows that the scheme is both robust in additive white Gaussian noise channel and slight non-line-of-sight channel scenarios. The performance in multipath scenarios has also demonstrated the reliability and feasibility of our scheme.

Artificial bee colony (ABC) algorithm is a swarm intelligence based optimization technique, which has attracted wide attention from different research fields. In the basic ABC, however, the same solution search equation is used in both of the employed bee phase and onlooker bee phase, which performs well in exploration but poorly in exploitation. To address this concerning defect, in this paper, we propose an improved ABC variant by designing a mechanism of utilizing directional information. In this mechanism, we first construct a pool of differential vectors in the employed bee phase, and then utilize a differential vector randomly selected from the pool as directional information to guide search in the onlooker bee phase. Furthermore, we propose two novel solution search equations based on the current best solution and some good solutions with the aim of balancing exploration and exploitation. Experiments are conducted on a set of 22 well-known benchmark functions, and the results demonstrate that our proposed approach shows promising performance.

As an effective swarm intelligence based optimization technique, artificial bee colony (ABC) algorithm has become popular in recent years. However, its performance is still not satisfied in solving some complex optimization problems. The main reason is that both of the employed bee phase and onlooker bee phase use the same solution search equation to generate new candidate solutions, and the solution search equation is good at exploration but poor at exploitation. To solve this problem, in this paper, we propose a multi-strategy artificial bee colony algorithm with neighborhood search (MSABC-NS). In MSABC-NS, a multi-strategy mechanism is designed to use two different solution search equations, and a neighborhood search mechanism is introduced to make full use of good solutions. Experiments are conducted on 22 widely used benchmark functions, and three different ABC variants are included in the comparison. The results show that our approach can achieve better performance on most of the benchmark functions.

With the fast integrative development of the Internet, the Internet of things, and mobile communications, it has become an interesting issue to study human’s psychology and behaviors from the big data of the interrelated spaces based on machine learning. This article analyzes the new requirements of machine learning for CPP (Cyber Psychosocial and Physical Computation), and presents the concept, framework, method, and application of smart learning, which refers to the comprehensive learning ability of rational intelligence and emotional intelligence. In order to better understand the relationships among different spaces and reduce the computation cost by prior-knowledge, the authors propose the above learning process to be built up on the basis of neural mechanism and context awareness. Research of this article aims to provide a reference for the development of machine learning to deal with human’s psychology and behaviors under the ubiquitous environment.

The design of interactive art installation, which integrates interactivity, artistry and education, leads people to think their relationship with nature and society by human-installation interactive movement. Due to the repaid development of science and technology, the simple interaction with settled program cannot satisfy people’s demands. Instead, they want to step into the world of installation through personalized emotional catharsis. The current mode of communication will be changed by the in-application program within the interactive art installation. The single installation totally produced by artists’ thought will be replaced by the finished installation product with the participation of the audience. In this way, participants will immerse in the story plot and feel various emotional changes psychologically by sound, image and touch, which relies on the interactive art installation. The Listen is taken as the example to spread the idea of sharing the ecological nature of human beings and the world. Arduino, an open source platform, will be studied to show the change of modern interactive communication mode, exploring the impact of programming tool on the development of future interactive installation and providing some new ideas for the integration of artistic creation and science and technology in the future.

Artificial bee colony (ABC) is a type of popular swarm intelligence optimization algorithm. It is widely concerned because of its easy implementation, few parameters and strong global search ability. However, there are some limitations for ABC, such as weak exploitation ability and slow convergence. In this paper, a novel ABC with adaptive neighborhood search and Gaussian perturbation (called ABCNG) is proposed to overcome these shortcomings. Firstly, an adaptive method is used to dynamically adjust the neighborhood size. Then, a modified global best solution guided search strategy is constructed based on the neighborhood structure. Finally, a new Gaussian perturbation with evolutionary rate is designed to evolve the unchanged solutions at each iteration. Performance of ABCNG is tested on two benchmark sets and compared with some excellent ABC variants. Results show ABCNG is more competitive than six other ABCs.

Accident causation theories explain the possible causation mechanisms of accidents, and unsafe behavior is a major component of such causes. Considering the limitation that existing studies rarely discuss the interactions among unsafe behaviors in accidents, this paper uses a case study in the Chinese building construction industry to explore the behavioral risk chains of accidents based on complex network (CN) theory. First, accident cases are collected from government websites, and the extracted unsafe acts are classified according to a list summarized based on several safety standards and operating procedures. Second, the rules for forming behavioral risk chains are defined, and then, a behavioral risk chain network of accidents (BRCNA) is established. In addition, Pajek is used to construct the network model. Finally, the topological parameters are calculated and analyzed in the BRCNA. The results show that the BRCNA has the properties of a scale-free and small-world network. These findings indicate the robustness of the BRCNA for random attacks and the high transmission and diffusion efficiency of behavioral risk in the BRCNA, which reflects that some unsafe acts must be critically controlled and that their related unsafe acts in a chain must be collaboratively controlled through safety management. This study is of theoretical and practical significance for accident prevention in the construction industry.

A Gaussian multiple-input single-output (MISO) fading channel is considered. We assume that the transmitter, in addition to the statistics of all channel gains, is aware instantaneously of a noisy version of the channel to the legitimate receiver. On the other hand, the legitimate receiver is aware instantaneously of its channel to the transmitter, whereas the eavesdropper instantaneously knows all channel gains. We evaluate an achievable rate using a Gaussian input without indexing an auxiliary random variable. A sufficient condition for beamforming to be optimal is provided. When the number of transmit antennas is large, beamforming also turns out to be optimal. In this case, the maximum achievable rate can be expressed in a simple closed form and scales with the logarithm of the number of transmit antennas. Furthermore, in the case when a noisy estimate of the eavesdropper's channel is also available at the transmitter, we introduce the SNR difference and the SNR ratio criterions and derive the related optimal transmission strategies and the corresponding achievable rates.

Firefly algorithm (FA) is a new swarm intelligence optimization method, which has shown good search abilities on many optimization problems. However, the performance of FA highly depends on its control parameters. In this paper, we investigate the control parameters of FA, and propose a modified FA called FA with adaptive control parameters (ApFA). To verify the performance of ApFA, experiments are conducted on a set of well-known benchmark problems. Results show that the ApFA outperforms the standard FA and five other recently proposed FA variants.

As a relatively new optimization technique, in recent years, artificial bee colony (ABC) algorithm has attracted much attention for its good performance. However, its performance has also been challenged in solving complex optimization problems. This insufficiency is mainly caused by its solution search equation, which does well in exploration but badly in exploitation. Inspired by the concept of neighborhood search, in this paper, we introduce a global neighborhood search operator into ABC for balancing its explorative and exploitative capabilities. Extensive experiments are conducted on 22 benchmark functions, and six different algorithms are included in the comparison studies, including four ABC variants and two related evolutionary algorithms. The compared results demonstrate that in most cases our approach is able to provide better performance in terms of solution accuracy and convergence speed.

In the whole-core heterogeneous transport calculations with massive parallel, the computational cost of the Coarse Mesh Finite Difference (CMFD) method might become the most time-consuming part if directly using the traditional CMFD method. To accelerate the CMFD calculations, a new three-level CMFD method based on the loosely coupled parallel strategy is developed in the high-fidelity neutronics code NECP-X. Two aspects are considered to accelerate the CMFD calculations, one is reducing the CMFD matrix size and another is minimizing the number of outer iteration numbers. The three-level CMFD acceleration method includes: (1) the multi-group pin-level CMFD (MG-Pin-CMFD) for accelerating transport calculations, in which the cross sections and currents are obtained from the 2D-1D transport calculation; (2) the one-group pin-level CMFD (1G-Pin-CMFD) for accelerating the MG-Pin-CMFD calculations, and the 1G-Pin-CMFD is established by the group-condensed cross sections, flux and currents from the MG-Pin-CMFD; (3) the one-group assembly-level CMFD (1G-ASY-CMFD) for accelerating 1G-Pin-CMFD, and the assembly-wise parameters are from the spatial homogenization of the 1G-Pin-CMFD calculations. The three-level CMFD method can reduce the outer iterations of CMFDs, such as MG-Pin-CMFD and 1G-Pin-CMFD. In every level of CMFD, the group-wised CMFD linear system is constructed based on the loosely coupled spatial domain decomposition, and the serial Generalized Minimum RESsidual (GMRES) method is implemented to solve the linear system instead of the parallel GMRES method to minimize the communications. A set of benchmark problems are tested and the results show good performance and good accuracy in transport calculations.