Table of Content

29 November 2023, Volume 0 Issue 11

Previous Issue    Next Issue

Joint Extraction Method of Entities and Relations Based on FGM and Pointer Annotation

LIU Yu-peng, GE Yan, DU Jun-wei, CHEN Zhuo

2023, 0(11):  1-5.  doi:10.3969/j.issn.1006-2475.2023.11.001

Asbtract ( 197 )   PDF (1192KB) ( 75 )  

References | Related Articles | Metrics

Abstract: Joint extraction of entities and relations is an important task of information extraction. The traditional entity relationship joint extraction method cannot solve the problem of overlapping triples well， because it models the relationship between entities as discrete types. In order to solve the problem that it is difficult to extract overlapping triples， this paper proposes a BERT-FGM model for entity relationship joint extraction， which combines FGM and pointer annotation. In this model， the relationship between entities is modeled as a function， and the robustness of the model is improved by incorporating FGM into the process of BERT training word vector. The model firstly extracts the subjects through the pointer annotation strategy， then fuses the subjects into a sentence vector as a new vector， and finally uses it to extract objects under a predefined relationship condition. Experiments are carried out on public dataset WebNLG， the experimental result shows that the F1 value of the model is 90.7%， it can effectively solve the problem of relationship triples overlapping.

Cross-language Multi-label Sentiment Classification Based on Stacked Denoising AutoEncoder

TANG Shi-qi, ZHOU Rui-ping, XIE Shi-bin, LIU Meng-chi, XIAO Wen,

2023, 0(11):  6-12.  doi:10.3969/j.issn.1006-2475.2023.11.002

Asbtract ( 102 )   PDF (1392KB) ( 63 )  

References | Related Articles | Metrics

Abstract: The multi-label sentiment classification task aims to deal with the problem that an instance may be associated with multiple sentiment labels. Most existing multi-label sentiment classification models were designed based on complete data，and their performance and sentiment were easily affected by the incompleteness of data itself. To address this problem，a cross-language multi-label sentiment classification model based on stacked denoising autoencoder is proposed， and a loss function is introduced to compensate for the loss caused by training. In this model， the word vectors are denoised by the stacked denoising autoencoder to construct the low-dimensional features of the original data. This reduces the noise interference in feature space and provides effective feature representation for downstream tasks. In the multi-label sentiment classification experiment of SemEval2018 three language datasets （English， Arabic and Spanish）， the micro_F1 score， macro_F1 score and jaccard indexes of the model on the test set are all improved. Macro_F1 is improved by about 0.82， 1.45 and 1.83 percentage points， respectively.


Collaborative Device-based Large-scale Offloading: A Bi-level Optimization Algorithm Fusing Divide-and-conquer and Greedy

YAN Yang, ZHAN Zi-jun, CAO Shao-hua

2023, 0(11):  13-21.  doi:10.3969/j.issn.1006-2475.2023.11.003

Asbtract ( 76 )   PDF (2607KB) ( 69 )  

References | Related Articles | Metrics

Abstract: With the rapid development of communication technology， the number of mobile devices is constantly increasing， which will also lead to frequent large-scale offloading scenarios. However， solving large-scale offloading problems in polynomial time remains a challenge. In this paper， we propose a bi-level optimization algorithm based on the cooperative computing network architecture， called DCGreedy， which fuses divide-and-conquer and greedy. This algorithm can efficiently solve the offloading strategy and resource allocation scheme of all tasks in polynomial time. It can effectively reduce the total energy consumption of the system while meeting all constraints. We evaluate the performance of DCGreedy based on the total number of tasks meeting deadlines， total system energy consumption， and algorithm runtime in a simulation scenario of at least 400 mobile devices. We conducted extensive experimental comparisons between DCGreedy and four other benchmark algorithms and found that in different scale offloading scenarios， the average total energy consumption of DCGreedy was 2.11% higher than the second ranked algorithm， while the algorithm’s running time was only 0.0049%. This fully confirms that DCGreedy effectively reduces the algorithm’s running time while optimizing system energy consumption.


Survey of Supply Chain Oriented Consensus Algorithms

CHAI Li, WANG Xiao, GONG Jia-hao, WANG Yang, JI Shun-hui, ZHANG Peng-cheng

2023, 0(11):  22-27.  doi:10.3969/j.issn.1006-2475.2023.11.004

Asbtract ( 119 )   PDF (993KB) ( 69 )  

References | Related Articles | Metrics

Abstract： As one of the core technologies in the blockchain， consensus algorithm is an important method for the system to maintain data consistency and security. This paper firstly investigates and analyzes the relevant research on the universal consensus algorithms in the alliance chain， classifies algorithms from the perspective of whether they are based on the Byzantine problem， and combs and summarizes consensus algorithms from four aspects： problem entry， principle description， performance analysis and application scenarios. In addition， focusing on the application scenarios related to the supply chain， this paper analyzes the challenges it brings to the consensus algorithm in the alliance chain， and sorts out and summarizes the consensus algorithms in the alliance chain under this scenario. Finally， the paper discusses the challenges faced by the consensus algorithm and the direction for future development， with an intention of providing references for researchers in this field.


Load Balancing Algorithm of Multi-job Cluster Based on SDN and Improved CSA Algorithm

WANG Chong-yang, ZHUANG Yi

2023, 0(11):  28-35.  doi:10.3969/j.issn.1006-2475.2023.11.005

Asbtract ( 50 )   PDF (1460KB) ( 59 )  

References | Related Articles | Metrics

Abstract： In order to achieve more efficient task scheduling， reduce the total task completion time of the system， ensure reliable service performance， and improve the stability， flexibility， and high availability of multi-job cluster systems， this paper proposes a method based on SDN and SOS-ICSA （symbiotic organisms search and improved crow search algorithm） load balancing strategy to improve the scheduling problem of the multi-job cluster management system. In order to improve the convergence speed of the CSA algorithm and the quality of the solution， a local optimization strategy is added to the CSA algorithm， and the SOS algorithm is used to solve the two problems of CSA. Adaptive optimization of each control parameter helps CSA find the best or near-optimal solution. The algorithm also takes into account the reliability of the virtual machine and combines with SDN to further improve the system performance and flexibility. The experimental results show that the algorithm proposed in this paper reduces the imbalance of the system， reduces the total execution time of tasks，and improves resource utilization.


Prediction of Diabetes Mellitus Using LightGBM Classifier with RF-RFECV

LIU Jing-le, LUO Xiang, GONG Cheng-rong, ZHANG Guo-peng

2023, 0(11):  36-43.  doi:10.3969/j.issn.1006-2475.2023.11.006

Asbtract ( 124 )   PDF (2220KB) ( 63 )  

References | Related Articles | Metrics

Abstract: In order to find the high-risk population of diabetes in China as early as possible and provide targeted intervention measures， the data set of China Health and Retirement Longitudinal Study （CHARLS）， which represents the Chinese population， was selected as the research object， and a hybrid algorithm based on RF-RFECV and LightGBM （RF-RFECV-LightGBM） was proposed， and compared with five other algorithms through experiments. The results show that RF-RFECV- LightGBM has the best overall performance， the accuracy， precision， recall， F1 value and AUC value are 0.9772， 0.9952， 0.8178， 0.8978， and 0.9357， respectively. The prediction time is 0.0428 s， which is 0.0549 s shorter than the prediction time of LightGBM before feature selection （increased by 56.19%）， indicating the effectiveness of RF-RFECV algorithm. Finally， the same prediction process was tested on the Pima Indian dataset， and the results achieved an accuracy of 0.9415， further verifying the excellent performance of the proposed algorithm RF-RFECV-LightGBM， which can assist in clinical diagnosis and treatment of diabetes.


Application Research of Implicit Feedback Recommendation Based on E-commerceUser Behavior

ZHU Hong-qi, WANG Cheng

2023, 0(11):  44-50.  doi:10.3969/j.issn.1006-2475.2023.11.007

Asbtract ( 70 )   PDF (911KB) ( 74 )  

References | Related Articles | Metrics

Abstract: The Bayesian Personalized Ranking algorithm is one of the most representative algorithms for implicit feedback problems， but both the assumption of independence between users and the assumption of individuals’ pairwise preferences for two items proposed in the BPR algorithm are too restrictive. The GBPR algorithm redefines the individual preferences of users， using group preferences formed by like-minded users instead of individual preferences to relax the assumption of independence among users. The DPR algorithm takes the partial order pair as the basic unit to optimize the difference between preferences rather than the difference of preferences to relax the assumption of an individual’s pairwise preference for two items. Based on the above research， this paper proposes an e-GDPR algorithm to further enhance the user’s ability to predict preferences for items. The algorithm can make full use of user information （such as gender， consumption level） and commodity information （such as category） in the data set， introduce group preference into the DPR algorithm， divide users into groups according to consumption level and gender， randomly sample to form more representative user groups， and no longer use random sampling directly when sampling.Instead， a triad sample consisting of two randomly selected commodities belonging to the same category is considered to be more reliable than a triad sample consisting of randomly selected commodities. Then the implicit feedback preference quantification model is introduced to calculate the user’s personal preference， which can fully consider the user’s preference behind various implicit operation types. Finally， a recommendation experiment is carried out on the Jingdong e-commerce data set， and the experimental results show that the e-GDPR algorithm can achieve better recommendation results compared with the baseline algorithm.



An Importance Assessment Method of Network Assets in Critical Information Infrastructure Based on Percolation Theory

HUANG Yu-ting, CHEN Lin, LIN Hong-gang,

2023, 0(11):  51-56.  doi:10.3969/j.issn.1006-2475.2023.11.008

Asbtract ( 50 )   PDF (1185KB) ( 53 )  

References | Related Articles | Metrics

Abstract: The assessment of the importance of network assets of critical information infrastructures is a key national concern at present. To address the problem that the current network asset importance assessment ignores the business chain and thus affects the accuracy and validity of the results， this paper constructs a three-layer coupled network of “information-physical-users” based on the supply-demand relationship of network services， and proposes an asset importance assessment method based on the network percolation theory: we apply the improved network percolation theory to the coupled model and describe the propagation of failure in the network by combining the percolation probability of nodes and the loss of resource delivery capacity of nodes， and then integrate the change rate of maximum service delivery load and user impact level of nodes before and after failure to distinguish the different impacts of nodes. Finally， simulation experiments are conducted in the context of the electric power industry， and the results show that the method in this paper has high accuracy and provides a theoretical basis for the importance assessment of network assets.


Analyzing to Shield Tunnel Segments Deformation Data Based on ICEEMDAN-LSTM

FENG Xin-xin, BU Lei, ZHANG Xiao-yu, SHI Yu-feng

2023, 0(11):  57-61.  doi:10.3969/j.issn.1006-2475.2023.11.009

Asbtract ( 56 )   PDF (3338KB) ( 69 )  

References | Related Articles | Metrics

Abstract：Measures of subway tunnel safety monitoring and monitoring data analysis and prediction are important means to ensure the safety of subway tunnel. Due to the influence of construction environment， there are noise in the monitoring data inevitably. Taking the automatic deformation monitoring data of shield subway tunnel segments as the research object， a deformation monitoring data analysis and prediction method was presented based on ICEEMDAN-LSTM. Firstly， ICEEMDAN was used to decompose the monitoring data and obtain the IMF and residual components of the monitoring data. The LSTM network model was built， and it was used to predict the IMF and residual components of the monitoring data. Finally， the predicted values of IMF and residual components were superimposed and reconstructed to obtain the predicted values of deformation. The experimental results show that ICEEMDAN-LSTM model has higher prediction accuracy than BP and LSTM model.


QoE-driven High-availability Transmission Framework for SDN

CHEN Chen, ZHUANG Yi, GAO Zeng

2023, 0(11):  62-68.  doi:10.3969/j.issn.1006-2475.2023.11.010

Asbtract ( 63 )   PDF (1769KB) ( 72 )  

References | Related Articles | Metrics

Abstract: For the problem of low availability and high cost of path switching for single-path transmission in complex environments， this paper proposes a QoE-driven and SDN-assisted MPTCP path switching scheme for high-availability transmission services. Firstly， a path planning model with path disjoint is constructed based on the mesoscopic centrality of data plane nodes in SDN architecture. Secondly， a two-stage strategy is used to perform bandwidth compensation and path switching separately， the transfer of subflows on the transmission blocked path is accelerated to achieve path switching with lower throughput. Finally， Levy flight is introduced into the model update of AOA to prevent the algorithm from converging prematurely and enhance the ability to jump out of the local optimum， thus ensuring that the algorithm is optimal when optimizing the subflow path weights. Experimental results show that the method proposed in this paper has the advantages of smaller fluctuation range of throughput， lower delay and less jitter when performing sub-stream path switching. In addition， the improved AOA algorithm can obtain higher convergence efficiency when calculating the optimal weight vector.



Multi-user Dynamic Bandwidth Allocation Method Based on Edge Computing

YANG Bo

2023, 0(11):  69-74.  doi:10.3969/j.issn.1006-2475.2023.11.011

Asbtract ( 64 )   PDF (1241KB) ( 61 )  

References | Related Articles | Metrics

Abstract: Multiple cluster frameworks in the same data cente increased the allocation time of dynamic bandwidth. In view of this situation， researchers proposed a multi-user dynamic bandwidth allocation method based on edge computing. This method uses a recurrent neural network to calculate and construct an allocation model， and applies elastic bandwidth sharing through a three-layer allocation mechanism and link framework. The priority flow level of user business is set through a sharing mechanism to promote the synchronous transmission of business needs among multiple users， while achieving dynamic bandwidth allocation among multiple users. Experimental results show that the new method can be classified based on different scale of users， and has high allocative efficiency and practicability.


Critical Fault Point Identification Method for Routing Algorithms Using XGBoost

LI Xiang, ZHUANG Yi

2023, 0(11):  75-81.  doi:10.3969/j.issn.1006-2475.2023.11.012

Asbtract ( 27 )   PDF (2171KB) ( 45 )  

References | Related Articles | Metrics

Abstract: It is particularly important to ensure the reliability of routing algorithms under the single event effects. To address the problem of excessive overhead in identifying program critical fault points by exhaustive fault injection， this paper proposes a critical fault point identification method for routing algorithms using XGBoost. The method firstly maps the unit flip caused by the single event effects into the program instructions of the routing algorithm and builds a fault model; then uses this fault model to guide the construction of the fault point feature vector and uses the XGBoost algorithm to train a fault point fault type prediction model； finally identifies the critical fault points in the routing algorithm based on the model prediction results. The experimental results show that， compared with other methods， the key fault point identification method of routing algorithm using XGBoost proposed in this paper has a higher identification rate and reduces the overhead caused by the exhaustive fault injection method.


Real-time Rendering Algorithm for Spatial Targets in Air-space 3D Simulation

ZHANG Chun-hui, NIE Yun, WANG Guo-wei

2023, 0(11):  82-88.  doi:10.3969/j.issn.1006-2475.2023.11.013

Asbtract ( 55 )   PDF (1894KB) ( 65 )  

References | Related Articles | Metrics

Abstract: In recent years， with the deepening research of manned spaceflight， the complexity and reliability requirements of space missions have been increasing. Real-time position calculation of massive targets and scene rendering are the key and difficult points of real-time rendering of space targets. To take advantage of the hierarchical detail model （LOD） in dynamic rendering， a real-time rendering method for massive space targets is proposed， which focuses on optimizing the traditional batch LOD model into an R-tree-based LOD model. When constructing the R-tree-based LOD model， there are issues such as index space overlap， low query efficiency， and LOD model texture mutation. Therefore， a node-based depth adjustment strategy is proposed to eliminate index space overlap， a fast pruning algorithm is adopted to improve query efficiency， and the shader-based alpha testing technique is used to achieve smooth transitions between LOD models. Through the collaborative processing of these three optimization algorithms， the optimized LOD model has improved in scene rendering time， space occupancy rate， frame rate， and other aspects.


Lightweight Facial Expression Recognition Method Based on Sandglass Structure and Attention Mechanism

LUO Ming-jie, FENG Kai-ping

2023, 0(11):  89-94.  doi:10.3969/j.issn.1006-2475.2023.11.014

Asbtract ( 64 )   PDF (1621KB) ( 74 )  

References | Related Articles | Metrics

Abstract: Facial expression detection and classification is a challenging task in the field of human-computer interaction. In order to solve the problems of large parameters and low classification accuracy in current facial expression recognition models， a lightweight facial expression recognition method based on sandglass structure and attention mechanism is proposed. First， the improved sandglass structure is used to build a lightweight backbone feature extraction network. Then a novel feature fusion attention module is designed. Focus pooled features are fused to extract key details， and lightweight ECA attention mechanism is embedded to strengthen key expression features to improve the feature expression ability of the model. Finally， various training strategies such as Random Erasing and Dropout are adopted to alleviate the over fitting phenomenon of lightweight networks， so as to improve the generalization performance of the model. Testing experiments were conducted on two classical expression datasets FER2013 and CK+， and the recognition rates reached 71.72% and 95.96% respectively， while the number of parameters is only about 1×106.


Real-time Detection of Arbitrary Shape Scene Text Based on Segmentation

XU Hong-kui, LI Zhen-ye, GUO Wen-tao, ZHAO Jing-zheng, GUO Xu-bin

2023, 0(11):  95-100.  doi:10.3969/j.issn.1006-2475.2023.11.015

Asbtract ( 125 )   PDF (1710KB) ( 68 )  

References | Related Articles | Metrics

Abstract：The current challenges of scene text detection technology are mainly reflected in two aspects: the trade-off between model real-time performance and accuracy， and the detection of arbitrary shape text. They determine whether scene text detection is feasible in real scenes. Aiming at the above two problems， this paper proposes a lightweight backbone network with strong feature extraction ability based on segmentation method， which can accurately detect natural scene text of arbitrary shape in real time. Specifically， a simple dual-resolution residual backbone network and a deep aggregate pyramid pooling module with low computational cost are used， and the features extracted from them are fused and segmented using a differentiable binarization module. Through the comparative experiment on the standard English dataset ICDAR2015， the result show that the improved method proposed in this paper is effective， and achieves comparable results in real-time performance and accuracy.


Behavior Recognition Method Based on FMCW Radar and ResNeSt-GRU

MA Ze-yu, YE Ning, XU Kang, WANG Su, WANG Ru-chuan,

2023, 0(11):  101-107.  doi:10.3969/j.issn.1006-2475.2023.11.016

Asbtract ( 68 )   PDF (2574KB) ( 76 )  

References | Related Articles | Metrics

Abstract： Aiming at the application of frequency modulated continuous wave radar in behavior recognition， a human behavior recognition system based on split attention residual neural network （ResNeSt） and gated neural unit （GRU） is proposed. The frequency modulated continuous wave （FMCW） radar is used to collect human behavior data. The fast Fourier transform algorithm （FFT） is used to extract the distance， velocity and angle dimension information of each frame of radar data， and then stitch them according to the time dimension into Range-Time Map （RTM）， Doppler-Time Map （DTM） and Angle-Time Map （ATM）. Finally， RTM， DTM and ATM are used as input samples， and the three-stream ResNeSt-GRU model is used to recognize different human behaviors. The experimental results show that the average recognition accuracy of the three-stream ResNeSt-GRU model for 8 behaviors reaches 98.92%， which is higher than the traditional deep learning model and the fusion deep learning model. In addition， the recognition accuracy rate using this model is 2.3% higher than that using a single-stream network after traditional feature fusion. Therefore， the system can effectively improve the recognition accuracy of the human behavior recognition system， and provide a new technology for the human behavior recognition.


REST Web Service System for Image Data Based on Beidou Grid

LIU Fu, YU Jin-song-di, WEI Dan-dan,

2023, 0(11):  108-112.  doi:10.3969/j.issn.1006-2475.2023.11.017

Asbtract ( 59 )   PDF (1434KB) ( 61 )  

References | Related Articles | Metrics

Abstract: Multi-source and heterogeneous Earth Observation （EO） data is widely used in different fields of national economic construction. However， different spatial reference frames have been increasingly adopted for data of different types， sources and sectors， which pose a big challenge to the sharing of spatial information of different reference frames. Promisingly， the combination of global subdivision grid and Web services provides a fine solution to such a problem. Among them， the Beidou grid has the characteristics of simple segmentation structure and easy to use， which can assist in retrieving image data. Web services can be used for image data retrieval under the Beidou grid code identification. Therefore， this paper takes the Beidou grid as a reference， establishes the mapping relationship between grid and image data， proposes a Beidou grid description method and image data acquisition method for remote sensing image data， and designs a prototype of Beidou grid image description Web service and Beidou grid code based image data acquisition Web service based on RESTful architecture， achieving interoperability of image data using grid as a unit under the same reference framework.


CP-YOLOX-based Algorithm for Protein Target Detection in Cryo-electron Micrographs

OU Jia-cheng, ZENG An, JIN Liang

2023, 0(11):  113-119.  doi:10.3969/j.issn.1006-2475.2023.11.018

Asbtract ( 52 )   PDF (2886KB) ( 80 )  

References | Related Articles | Metrics

Abstract： A cryo-electron micrograph target detection algorithm （Cryo-Protein YOLOX， CP-YOLOX） is proposed for the existing cryo-electron micrograph protein target detection algorithm with inadequate feature fusion and complex network model， missed detection and false detection. The algorithm mainly contains feature extraction module， feature fusion module， and output side. The feature extraction module applies the B-ResBlockX module proposed in this paper， which uses grouped filters to generate multiple feature channels to improve the feature fusion capability and capture more detailed features. The feature fusion module applies the FastHead module proposed in this paper， which uses multilevel dilated convolution module for feature fusion and simplifies the output to a single channel， which can have a more lightweight network structure without losing accuracy. In order to further improve the accuracy and convergence speed， the position loss function is added with the Euclidean distance constraint between the target frame and the prediction frame. Experimental results on public datasets EMPIAR-10028， EMPIAR-10081， and EMPIAR-10089 showed that the number of network parameters of the proposed algorithm was only 5.19×106， and the mAP（0.5） was improved by 2.4， 3.3 and 2.5 percentage points， respectively， compared with YOLOX.


Improved YOLOv7 Algorithm for Low-resolution Ship Object Detection in Complex Backgrounds#br#

YAN Zi-xian, DONG Bao-liang, TANG Si-mi

2023, 0(11):  120-126.  doi:10.3969/j.issn.1006-2475.2023.11.019

Asbtract ( 144 )   PDF (3631KB) ( 85 )  

References | Related Articles | Metrics

Abstract: In response to the problems of low resolution target detection and interference from complex backgrounds in ship image target detection， an improved YOLOv7 algorithm is proposed for identifying ship targets. The algorithm is mainly improved in three aspects: using K-means++ algorithm for anchor box clustering in the ship target dataset to obtain anchor box information that is more suitable for ship detection tasks; improving the loss function by using EIOU loss instead of CIOU loss and using Focal loss combined with ɑ-Balanced instead of standard cross-entropy loss; improving the network structure by adding the SPD-Conv module to enhance the detection effect for low-resolution targets. Experimental results show that compared with the original YOLOv7 algorithm， the improved algorithm has an accuracy improvement of 4.22 percentage points， a recall rate improvement of 2.68 percentage points， a mAP@0.5 improvement of 4.3 percentage points， and a detection speed improvement of 2 frames/s. The algorithm achieves good detection results for ship targets.