Search

Copyright of Chinese Journal of Liquid Crystal & Displays is the property of Chinese Journal of Liquid Crystal & Displays and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In order to improve the efficiency of particle-based fluid surface reconstruction, this paper proposed a perspective-grid-based adaptive narrow-band surface particle extraction method. Compared with the object-space-based method, this scheme adaptively extracted the surface particles closest to the viewpoint within the frustum according to the particle density, dispersion coefficient, so that the number of surface particles and memory consumption were related to the visible surface area, rather than the entire fluid surface or simulation domain. In addition, this paper presented an adaptive thickness estimation method based on particle density by taking advantage of the arrangement of perspective grids along the view. Experiments show that this scheme effectively reduces surface particles by 40% to 76% and memory overhead by 30% to 50%,solves the problems of surface particle redundancy and holes, and quickly obtains thickness information at a lower cost. This scheme brings significant performance improvement for surface reconstruction and rendering, and can better handle the surface reconstruction and rendering of large-scale particle sets. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Architectural Journal / Jian Zhu Xue Bao is the property of Architectural Journal Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Frontiers of Computer Science & Technology is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Under a large-scale and complex environment, for the path planning algorithm of JPS, the memory consumption is large, the smoothness of path result is low, and the path result is excessively close to obstacles. Aiming at these problems, this paper proposed an improved JPS algorithm that integrated a level function of safety potential field and an optimized Floyd algorithm. It firstly built the safety level function and reassigned the grid status in the grid map to build a safety level map. Then by introducing two bias functions of goal and cardinal direction, and by combining the safety level function, this paper further reduced the time consumption brought by symmetric search, which improved the safety level of the path planned. Besides, it also optimized Floyd algorithm by adding a flow of second smoothing algorithm. At last, the paper integrated B-spline interpolation, improving the smoothness of path planned by the improved algorithm. The simulation tests verify that there are significant improvements in memory consumption, path length, path smoothness and path safety level for the improved and optimized algorithm. [ABSTRACT FROM AUTHOR]

Copyright of Society: Chinese Journal of Sociology / Shehui is the property of Society: Chinese Journal of Sociology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Anomaly detection aims to capture the discriminative features with limited training samples. However, when some anomalies share common compositional patterns with the normal training data, the model likely reconstructs the anomalies well, leading to the miss detection of anomalies. To mitigate this drawback, this paper proposed a novel contrastive memory network, which used the contrast learning framework to separate the sample features based on the autoencoder, and then designed a memory network to suppress the normal features similar to anomaly. This method proposed a two-stage framework for detecting abnormal events. In the first stage, the method used contrastive learning to increase the difference between normal features and abnormal features, and gained representation to be the augment memory and suppression memory of memory network. In the second stage, the model used augment memory to record multi-time normal behavior features, and used suppression memory to constrain the expression of pseudo anomaly items in the augment memory. The AUC value reached 83. 26% on UCF Crime datasets and 87.53% on Shanghai Tech datasets, which were 1. 14% and 2.43% higher than the existing methods. The results demonstrate that this method can efficiently predict the temporal localization of anomaly events. [ABSTRACT FROM AUTHOR]

Copyright of Telecommunication Engineering is the property of Telecommunication Engineering and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Chinese Journal of Wildlife / Yesheng Dongwu Xuebao is the property of Chinese Journal of Wildlife Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aiming at the common reliability problems in the oral history works of architecture and urban planning, based on the dialectical understanding of the reliability characteristics of various historical materials, this paper analyzes the causes of reliability problems from three aspects: some inherent characteristics of human memory, the organization of oral history work, and the sorting and processing of oral records. It discusses the reasons for the reliability problem, proposes countermeasures for the research of architecture and urban planning history from the perspective of mutual verification of various historical materials, and puts forward a number of suggestions for the oral history of architecture and urban planning.. [ABSTRACT FROM AUTHOR]

Copyright of International Architecture is the property of Omniscient Pte. Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Application Research of Computers / Jisuanji Yingyong Yanjiu is the property of Application Research of Computers Edition and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Objectives: In recent years, the brain‑like navigation algorithm is a new research hotspot, which is expected to achieve autonomous navigation by imitating the ability of biological navigation. The core is‑ sue is how to improve generalization ability. Methods: This paper introduces the research background and theoretical basis of the brain‑like navigation algorithm. After investigation, we propose a computational framework of brain ‑ like navigation algorithm. The outstanding works in this field are discussed and analyzed under this framework, and we carried out experimental verification of some basic methods. Results： The main contributions of this paper are: (1) Comprehensively introduces and summarizes the theoretical basis and outstanding works in this field. (2) Proposes the computational framework of the brain‑like navigation algorithm, which scientifically defines the functions of different modules of the algorithm. (3) Through theoretical analysis and experimental verification, we summarized valuable conclusions and expectations. Conclusions: In terms of model design, mature methods of deep learning can also be applied to this problem, but need more modifications to further improve navigation capabilities; in terms of model training, combining the advantages of multiple learning algorithms is hopeful to further improve the generalization ability. [ABSTRACT FROM AUTHOR]

In this paper, we explore how factivity-related cognitive attitude verbs in Chinese can be categorized. Verbs such as zhidao 'know' and mingbai 'understand' are factive verbs in contrast with renwei 'think' and yiwei 'think (wrongly)', which are non-factive verbs, in Chinese. Surprisingly, jide 'remember' turns out to be factivity-alternating. We consider Chinese in comparison with the English type and the type of Altaic languages which include Korean. It is known that Altaic languages show factivity alternation of such verbs as 'know,' 'remember,' etc, between factive vs. non-factive (Lee, 1978; 2020), whereas English is regarded as largely a non-alternating language (except in negative or nonveridical contexts). Chinese, on the other hand, rather strictly reveals a typical type of non-alternating language: zhidao 'know' is constantly factive and has no non-factive alternant; renwei 'think' is typically non-factive. But jide 'remember' reveals factivity alternation with factive vs. non-factive readings, with no syntactic differences in embedded complement contents. Some scholars have proposed non-facitve alternants of zhidao and jide, and raised such a view as the so-called 'factivity variation/shift.' However, all of the non-factive alternants in their arguments turn out to be usages of modals or the tendency of 'implicative verbs' (Karttunen, 1971), which involve volitional modality (their complement subjects are co-referential with their main clause subjects, the verbs being action verbs). [ABSTRACT FROM AUTHOR]

The index method of in-memory key value storage determines the time performance and space overhead of the system, which is a key factor for improvement and optimization. Hash index provides access operations in O（1） time complexity, but it generates storage conflicts and causes access performance degradation. To this end, this paper proposed a bitmap-based hash optimization method for key value storage, which can avoid storage conflicts and improve access performance. This method divided the keys with common prefixes into the same block to reduce the storage space of keys, used a hierarchical bitmap structure in the block, where the global bitmap represented the suffix part of all keys to avoid storage conflicts, and the summary bitmap supported quick positioning and acceleration for range queries. The experimental results show that the optimized hash index can achieve high throughput and good concurrency performance on various workloads, and the memory consumption is greatly reduced compared with existing schemes. [ABSTRACT FROM AUTHOR]

CRAFT is a lightweight tweakable block cipher proposed in FSE 2019, which is suitable for the resource-constrained equipment with the small hardware area to protect information security. It adopts the128 bit key and 64 bit tweak to encrypt a 64 bit plaintext. The evaluation of its safety can provide a theoretical basis for future use. The paper researched the structural properties of CRAFT and the redundancy of its key schedule, and chose a 13-round impossible differential of CRAFT, which made full use of the linear relations of round subkeys. Appended three rounds and five rounds at its top and bottom, this paper proposed a 21-round impossible differential, which applied some techniques such as the precomputation tables, the differential enumeration and the key-dependent property. The time, data and memory complexities of the attack were 296.74 encryptions, 253.6 chosen plaintexts and 256.6 64 bit blocks, respectively. This impossible differential attack on reduced-round CRAFT was the lowest time complexity attack under the single-key and single-tweak model. This method relied on the linear correlation of the handle value tweakey schedule, which was helpful to further understand the design of CRAFT. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Beijing University of Technology is the property of Journal of Beijing University of Technology, Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The emerging non-volatile memory ( NVM) is byte-addressable, which has the characteristics of low latency similar to memory and the nonvolatile property of storage. Limited by the maturity of hardware and software technology, NVM was firstly used in storage. This paper discussed a series of problems and management challenges of NVM for storage. Then, it summarized the main features of the existing typical NVM file systems. These features mainly focused on reducing consistency overhead, reducing the software stack cost, the integrated management of memory and storage, distributed file system, NVM file system security, fault tolerance, space management and so on. Finally, this paper prospected several research directions of NVM file system, including scalability, fusion of virtual memory and file system, and distributed file system. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Chongqing University of Technology (Natural Science) is the property of Chongqing University of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Background: The author of this paper works in the critical care ward of an internal medicine department. During the study period, current trainees in a nurse post graduate year (NPGY) program joined our nurse team. While at our ward, they were required to demonstrate clinical nursing competencies in intensive care units in the facets of nursing knowledge, equipment use, prediction of emergencies, emergency response, and patient evaluation and management. The ability to interpret arrhythmias is critical to connecting and coordinating these competencies. However, arrhythmia interpretation is usually the most difficult skill for nurses to master. Purpose: This study was developed to improve the knowledge of NPGY trainees regarding arrhythmia interpretation, the accuracy of their arrhythmia interpretation and management, and their confidence in caring for patients with arrhythmia. Resolution: The period of this study spanned from April 5th to June 18th, 2018. After surveying the current capabilities of the NPGY trainees to provide a reference for improvement, several teaching strategies were adopted. These strategies included: (1) provision of arrhythmia clinical scenarios; (2) establishment of multimedia teaching and interactive e-books; (3) development of memory strategies; and (4) mind mapping. Results: The accuracy of the trainees' knowledge regarding arrhythmia interpretation increased from 52.5% pretest to 92.5% posttest, while their arrhythmia interpretation skill improved from 56.3% pretest to 92.5% posttest. Furthermore, their caring-for-patients-with-arrhythmia confidence score increased from 5.5 to 9.1. All of the objectives of this study were achieved. Conclusions: The diverse teaching approach employed in this study enhanced the ability of trainees to remember and recall relevant theories as well as improved their related practical skills. Furthermore, posttest, the trainees were significantly less nervous during encounters with patients with arrhythmia and demonstrated improved problem-solving abilities. Participation in this intervention significantly improved the confidence of trainees to serve as critical care nurses and to pursue a career in critical care nursing. [ABSTRACT FROM AUTHOR]

Copyright of Electric Machines & Control / Dianji Yu Kongzhi Xuebao is the property of Electric Machines & Control and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Most existing class incremental learning methods employ data storage or extended network structures, but they cannot effectively alleviate the catastrophic forgetting problem due to memory resource limitation. To solve this issue, this paper proposed a brain-inspired generative replay approach. Firstly, it used VAE-ACGAN to simulate the memory self-organizing system to improve the quality of the generated pseudo-samples. Then, it used a shared parameter module and a private parameter module to protect the extracted features. Finally, the potential variables in the generator used a Gaussian mixture model to select specific replay pseudo-samples. The experimental results of the proposed method on MNIST, Permuted MNIST and CIFAR-10 show that the classification accuracy of the proposed method is 92.91%, 91.44% and 40.58%, respectively, which is significantly better than other class incremental learning methods. Furthermore, the backward transfer and forward transfer metrics reach 3.32% and 0.83% on the MNIST dataset, demonstrating that the model achieves a trade-off between task stability and plasticity, effectively preventing catastrophic forgetting. [ABSTRACT FROM AUTHOR]

Memory network only focusing on attention mechanism can't tackle this situation where the context sentiment relies on the concrete target. In order to address this problem, this paper proposed a target-sensitive deep memory network model. This model exploited attention mechanism to capture the sentiment information of context corresponding to a given target. Then the interaction module integrated context sentiment expression and interaction information between context and target to the classification features. Finally, it classified the features to get sentiment polarity of target. This paper conducted experiments on the two datasets from SemEval 2014 task4. The proposed model achieved better performance than attention-based memory networks. The experimental results show that considering the interaction between context and target is effective to solve the problem that context sentiment relies on the concrete target. [ABSTRACT FROM AUTHOR]

Password has always been the one of the most important way for identity authentication. To meet the requirements of password policy and memory, users often combine personal information as their passwords, therefore, in order to investigate the proportion of such passwords. This paper used 4 real password sets leaked in 2011 as experimental materials, preset the combination structure and format of the password and used an application to calculate the proportion of the password combined by personal information. Experimental results show that the proportion of pas swords combined with names, phone numbers, special dates and other information is about 12.41% to 25.53% . According to this rule, it proposed dynamic dictionary attack. An attacker can generate a dynamic dictionary and use it to crack the user password, after it obtains some of the users personal information. At last, this paper discussed that users how to select their passwords to prevent an attacker from cracking their passwords through a dynamic dictionary. [ABSTRACT FROM AUTHOR]

Traditional runtimes based on user estimating is usually less accurate. This paper combined the categorization with the instance-based learning method, used the template similarity and numerical similarity method to find the similar jobs of the current jobs in historical data, and used historical scheduling data to predict the runtimes of the current jobs. This paper only took seven job attributes into account, which included user name, group name, queue name, application name, requested number of processors, requested runtime, requested memory. It applied genetic algorithm to train the best parameters, and used similar jobs attributes to predict runtimes. Compared with the existing method, experimental results show that the proposed prediction method achieves a similar underestimate rate on the premise of using fewer parameters, and gets a lower mean absolute error. Moreover, on the HPC2N04 and HPC2N05 datasets, the mean absolute errors reduce 43% and 77% respectively. This paper studied the effect of using online prediction to replace user estimation on job scheduling, analyzed the results and pointed out the future improvement directions. [ABSTRACT FROM AUTHOR]

In order to understand the latest research progress of the classical clustering algorithm K-means in Spark environment, and grasp the current research hotspots and directions of K-means algorithm, this paper reviewed the initial center point optimization research on K-means algorithm. Firstly, it introduced the memory computing framework Spark and K-means algorithms, and analyzed the cause and effects of clustering instability of K-means algorithm, which pointed out the importance of optimizing K-means algorithm. This paper introduced the main methods and the latest research status of optimizing the initial center point of K-means in Spark environment in detail, and also discussed the future research trends in initial center point optimization of K-means. [ABSTRACT FROM AUTHOR]

In order to solve the shortcomings of the traditional serial encryption method in single machine mode, such as the low density of data security, the difficulty in encrypting-efficiency to meet the requirement of increasing information data encryption and the excessive amount of memory in the running process, and so on. This paper designed a scheme of chaotic encryption algorithm based on Hadoop platform. The scheme used MapReduce parallel framework, chaotic encryption pseudorandom number and the principle of initial value sensitivity. And it proposed parallel chaotic encryption scheme for MapReduce framework and chaotic encryption optimization. It used the length of the plaintext as the initial value, and used Chen, Lorenz, Rossler chaotic system to generate secret key. At the same time, the paper proposed an idea of dividing plaintext data into blocks according to 1 MB. According to the offset, it judged each block to gene rate three key sequences of Chenn Lorenzi R0ssler whose length is 1 MB, to improve the security of data density, and reduce the amount of running memory and other purposes. This design framework used Chen sequence for plaintext messing operation, used Lorenz sequence for XOR diffusion ope ration and used Rössler sequence for modular auxiliary diffusion operation. Experiments prove that, for MapReduce, the optimization algorithm of using parallel framework and chaotic system can effectively reduce the amount of memory and improve the encryption speed, and the encryption ope ration associated with plaintext achieves the purpose of effectively defending against the attack of selected plaintext. [ABSTRACT FROM AUTHOR]

The low level virtual machine obfuscator(OLLVM) is a well-known code obfuscation tool, which is not only used to protect the security of commercial software, but also used by malicious code developers to increase the difficulty of analysis. In order to facilitate the analysis of ARM malware by security researchers, this paper proposed and implemented an OLLVM automatic deobfuscation method based on dynamic analysis. For bogus control flow, this method monitored memory read and write based on the memory characteristics of opaque predicates and used dynamic taint analysis technology to identify it to complete deobfuscation; for control flow flattening, the method completed deobfuscation by dynamically running the program and recording the execution order of basic blocks; at the same time, multiple execution paths exploration was used to improve code coverage, and finally the relationship between basic blocks was restored through instruction repair. The experimental results show that the method could accurately eliminate the conditional branches caused by obfuscation in executable programs, and the running results of the deobfuscated programs was consistent with the unobfuscated programs. It verifies that the method can effectively complete the deobfuscation of ARM obfuscated programs. [ABSTRACT FROM AUTHOR]

Copyright of Foreign Literature Studies is the property of Foreign Literature Studies and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Piezoelectrics & Acoustooptics is the property of Piezoelectric & Acoustooptic and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In order to solve the problems of large number of parameters and loss of image texture and color when real images are transformed into animation style images, this paper proposed a multi-channel cartoon generative adversarial networks (MC＿CartoonGAN). Firstly, it reconstructed the multi-channel image data set by using HSCNN + and genetic algorithm, and obtained more abundant image information. Secondly, it used DenseNet for feature reuse to reduce the memory consumption of parameters and alleviate the problem of gradient disappearance. Finally, it reconstructed the loss function by introducing multi-channel color to reduce the loss of the generated image while ensuring the integrity of the generated image content. Experimental results show that the proposed multi-channel cartoon generative adversarial network(MC＿ CartoonGAN) has better quality in converting real images into cartoon style images. [ABSTRACT FROM AUTHOR]

In order to improve the correctness and efficiency of tree species recognition, to slow down the occurrence of overfitting, and to enhance the practicality of tree recognition techniques, this paper proposed a tree recognition method based on an improved convolutional neural network. The approach started with compressing the channel in the Xception framework to further exploit the average pool of global feature mapping, and changing the hybrid attention connection to parallel connection. Next, a randomly selected part of the feature map was normalized and re-entered into the neural network after using the attentional feature map cropping method. Finally, ablation experiment was performed and the accuracy of tree species recognition was up to 98.90% in the tree dataset when the learning rate was 0.1 and the iteration was 50 times. The study showed that the proposed improved convolutional neural network had better recognition effect on tree recognition. And the memory of the convolutional neural network architecture was reduced to 133.9 MB and the time consumed was only 458 ms. Using the improved convolutional neural network not only improved the accuracy of tree recognition, but also reduced the time cost. [ABSTRACT FROM AUTHOR]

To address the problem of large number of nodes and slow convergence rate of the RRT* algorithms in complex environments such as narrow passages, this paper proposed an improved efficient RRT* path planning algorithm, which based on multiple heuristic strategies and node enhancement mechanisms. It firstly adopted a combinatorial heuristic sampling strategy, by which the algorithm completed sampling within an informed ellipses subset after finding the initial path. It also introduced a heuristic node rejection strategy, which rejected nodes those did not meet the predefined conditions to participate in subsequent expansion. Moreover, the algorithm proposed a node enhancement mechanism to improve the utilization of nodes by enlarging the information contained in the nodes. The simulation results in complex environment with various obstacle distributions and narrow passages show that the improved algorithm HNEI-RRT* has fewer nodes and takes less time to converge than the other five improved RRT* algorithms, which verifies that the proposed algorithm can save memory space and increase convergence rate efficiently. [ABSTRACT FROM AUTHOR]

Diagnosability is a key indicator for evaluating the reliability of multiprocessor systems. The t/k diagnosis strategy greatly improves the diagnosability of the system by allowing at most k fault-free processors to be misdiagnosed as faulty processors. Compared with t-diagnosability and t1/t1 -diagnosability,t/k-diagnosability can better reflect the fault modes of a practical system. The 3-ary n-cube is a significant and common network topology, which is used as the underlying network in the con· struction of many distributed memory multiprocessors. According to some lemmas and the sufficient condition, this paper deter· mined that a system was t/k-diagnosable, when n ;?! 3 and 0 o⪯k o⪯ n, the 3-ary n-cube was tk,n/k -diagnosable, where tk,n/k = 2 ( k + 1) n - ( k + 1) ( k + 2) . This result shows that when choosing an appropriate value of k, the t/k-diagnos ability tk, n of 3-ary ncube is much greater than its t-diagnosability 2n and t1/t1 -diagnosability 4n - 3 [ABSTRACT FROM AUTHOR]

How to deploy neural networks in mobile or embedded devices with limited computing and storage capabilities is a problem that must be faced in the development of neural networks. In order to compress the model size and reduce the computational pressure, this paper proposed a neural network hybrid compression scheme based on the information bottleneck. Based on the information bottleneck, this scheme found redundant information between adjacent neural network layers and used this as a basis to trim redundant neurons, then performed ternary quantization on the remaining neurons to further reduce the model storage memory. The experimental results show that compared with similar algorithms on the MNIST and CIFAR-10 datasets, the proposed method has higher compression rate and lower calculation amount. [ABSTRACT FROM AUTHOR]

Aspect sentiment analysis is to analyze the emotional polarity of the target aspect item in a sentence. However, little research has been done on the influence of dependency between adjacent aspect items on sentiment classification. Based on this, this paper proposed an aspect sentiment classification model combining bidirectional LSTM based on attention mechanism and multi-hop end-to-end memory network for aspect sentiment analysis. Firstly, it used the sequential learning ability of Bi-LSTM, and introduced attention mechanism to obtain the semantic vector representation. Then, it used the multi-hop memory network to model the correlation between the target aspect and other aspects in the statement, and constructed the deep sentiment classification feature vector. Finally, it obtained the final sentiment polarity classification result by inputting it into softmax function. It tested this model on two datasets of restaurant and laptop in SemEval 2014 task and a set of public Twitter datasets, and improved the classification accuracy of the three datasets. The experimental results show that the effectiveness of the proposed model for aspect level sentiment classification and considering the dependency between aspects is beneficial for sentiment classification. [ABSTRACT FROM AUTHOR]

Recently, deep neural networks need to be deployed with low memory and computing resources, so it is necessary to design an efficient and compact network structure. This paper proposed a model compression method (KE ) based on improved attention transfer for the des ign of compact neural networks, which mainly used a wide residual teacher network (WRN) to guide a compact student network (KENet) by extracting both spatial and channel-wise attention to improve the performance, and applied this method to real-time object detection. The image classification experiment on CIFAR verifies that the knowledge distillation method with improved attention transfer can improve the performance of the compact model. The object detection experiment on VOC verifies that the model KEDet has good accuracy (72. 7 mAP) and time performance (86 fps) . The experimental results show that the object detection model based on improved attention transfer has good accuracy and real-time performance. [ABSTRACT FROM AUTHOR]

Copyright of Progress in Modern Biomedicine is the property of Publishing House of Progress in Modern Biomedicine and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Most of the existing dynamic RRT path planning algorithms cannot keep the planned path away from obstacles,which may cause the robot to have insufficient time to avoid obstacles.To solve this problem,this paper proposed an improved RRT,denoted as APFG-RRT,which utilized artificial potential fields to guide the RRT grow to goal and away from obstacles.In order to further increase the convergence rate and speed up the jump out of local minima,it introduced a strategy of selecting the goal as the random sample at an adaptive probability.Finally,it adopted global planning combined with local replanning to improve its real-time performance in dynamic environment.Simulation experiments show that APFG-RRT has higher computational efficiency and lower memory requirements compared with the initial RRT and Goal-bias RRT,and the given path can be effectively away from obstacles,which improves the success rate of dynamic path planning. [ABSTRACT FROM AUTHOR]

The nonlinear characteristics of RF power amplifiers is an important research direction. To address the problem of high sampling rate in the amplifier predistortion system, this paper proposed a new brodaband power amplifier predistortion model. Firstly, it used the modulation wideband converter ( MWC ) based on sinusoidal frequency modulation ( SFM ) signal to spare the signal in the feedback loop, and then used the variable step size generalized adaptive matching tracking( VS-GSA MP ) algorithm to reconstruct the signal. The experimental results show that normalized mean squared error ( NSME ) is about 2 - 3 dB better than memory polynomial( MP) model and the general memory polynomial( GMP) model. The adjacent channel power radio can be improved to 21 dB. Good linear performance can be achieved at a lower sampling rate. [ABSTRACT FROM AUTHOR]

Copyright of Telecommunication Engineering is the property of Telecommunication Engineering and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Virtualization technology is one of the important pillars of cloud services. Virtualization fully extends the flexibility of physical resources and improves the utilization of physical resources. With the development of national informatization level, the core technology of cloud servers is more and more independent, controllable, safe and efficient. In recent years, as a typical representative of domestic servers, Sunway has been fully developed. In order to further improve the performance of virtual machines in Sunway architecture, this paper proposes a method of extracting the memory access features of virtual machines in Sunway architecture. By combining the advantages of Sunway architecture, this pa per calculates the memory page miss ratio curve of virtual machines online based on the least recently used stack method, and reduces the performance cost of feature by using the hot set mechanism. Experiments show that the method can accurately calculate the working set size of virtual machine, the average error is less than 3%, and the average performance cost is no more than 8.3%. This work provides conditions for dynamic memory allocation of Sunway virtual machines, and finally improves the overall performance and memory utilization of Sunway cloud servers. [ABSTRACT FROM AUTHOR]

Due to the roughness of cache replacement mechanism in Spark parallel computing framework, LRU algorithm does not consider the RDD data block reuse problem that easily results in high-reuse block to swap out of the cache and execution inefficiency. By optimizing the weight model and improving the replacement strategy, this paper put forward ERAC cache strategy, which could achieve efficient RDD autonomous cache and hierarchical replacement, included high-reuse self-caching algorithm and cache replacement hierarchical algorithm. Applied ERAC, RA and LRU in standard dataset of SNAP, experimental results show that ERAC can effective improve the memory utilization and task execution efficiency. [ABSTRACT FROM AUTHOR]

The neural network structure with memory and attention mechanism shows some reasoning abilities needed for question answering. DMN is such an architecture, which achieves high accuracy in various language tasks. However, when there is no tag during training to support facts or whether it can be applied to other patterns such as images, does the architecture achieve strong results for question answering? Based on the analysis of DMN, this paper proposed several improvements of its memory and input module. Combining these changes, it introduced a new image input module to answer visual questions. The new dynamic memory model improved the latest technology on visual question answering data sets. The experimental results show that the proposed method achieves good results on visual question answering tasks. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Harbin Institute of Technology. Social Sciences Edition / Haerbin Gongye Daxue Xuebao. Shehui Kexue Ban is the property of Harbin Institute of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Chinese Studies / Hanxue Yanjiu is the property of Center for Chinese Studies/National Central Library and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

When training graph embedding via neural network, high dimension feature vector and large scale graph cause data transferring from memory to GPU to be a bottleneck. Aimed to solve this problem, this paper proposed graph partition based graph representation learning. This method splitted graph nodes into blocks according to their degree. It stored several node feature matrices in buffer pool on GPU. Every epoch, it trained representation during several blocks which fitted into buffer pool to reduce the data transferred from memory to GPU. Based on block split, this method used blocked based sampling algorithm, cached block feature matrix in GPU buffer pool to reduce memory read and built hierarchical negative sampling table, which could sample nodes in const time complex. Compared to related work on several real world datasets, this method achieves competitive accuracy at downstream machine learning task and 2 ~ 7 times speedup on training process. The experiments show that graph representation learning based on partition can train model efficiently and generate accurate embedding vectors. In future work, it is worth to prove the deviation between partition based method and original method in theory. [ABSTRACT FROM AUTHOR]