Search

In view of the problems of data sparsity and the diversity in academic paper recom-mender systems,based on CONVNCF, this paper proposed an algorithm of convolution with word and heterogeneous information network for academic paper recommendation ( WN -APR) . Firstly, WN -APR algorithm learned user and paper' s diverse features from different semantics to alleviate the sparsity problem. Then it designed an outer product fusing way to seamlessly combine user features with paper features. Replacing of 2D CNN, this algorithm applied 3 D convolution to mine the influence of different features on the performance. Finally, it modified the BPR loss function to enhance diversity in recommendations. Experimental results on CiteULike-a and CiteULike-t datasets show that WN-APR improves the performance of accuracy and diversity over the baseline models. [ABSTRACT FROM AUTHOR]

Most of the existing MAPF methods based on communication learning have poor scalability or aggregate too much redundant information, resulting in inefficient communication. To solve these problems, this paper proposed disruptor identifiable communication (DIC), which learned concise communication excluding non-disruptors by judging whether the agent in the center of the field of view would change its decision-making due to the presence of neighbors, and successfully filtered out redundant information. At the same time, this paper further instantiated DIC and developed a new highly scalable distributed MAPF solver: disruptor identifiable communication based on reinforcement and imitation learning algorithm (DICRIA). Firstly, the disruptor discriminator and the policy output layer of DICRIA identified the disruptor. Secondly, the algorithm updated the information of the disruptor and the communication wish sender in two rounds of communication respectively. Finally, DICRIA output the final policy according to the coding results of each module. Experimental results show that DICRIA S performance is better than other similar solvers in almost all environment settings, and the algorithm increases the success rate by 5.2% on average compared to the baseline solver. Especially in dense problem instances with large-size maps, the algorithm even increases the success rate of DICRIA by 44.5% compared to the baseline solver. [ABSTRACT FROM AUTHOR]

The estimation of multiple matching models between wide baseline or large angle images is a quite challenging task in image processing. The existing algorithms can be used to estimate multiple matching models and their inliers between images well, but their results are prone to matching pairs mis-classification issues. In order to accurately estimate the multiple matching models and allocate matching pairs, this paper proposed an image multi-matching model estimation algorithm based on the aggregation of matching points of the same model (AMPSM). Firstly, for improve the proportion of correct matching pairs, it filtered out incorrect matching pairs based on the distribution characteristics of correct matching points in the neighboring region. Furthermore, based on the different matching model degrees to which the matching pairs belong, searched for the suspected intersection matching pairs of multiple models, that was interference matching pairs. Meantime, for reducing the impact of interference matching pairs on the accuracy of matching classification, they were removed. Afterwards, for improve the clustering degree of matching points with the co-model, the position was dynamically moved based on the distance between the points within the same model and the center of gravity of the point set during the sampling process. Finally, classifying clustered matching points by Mean Shift to obtain a multi matching model. And the proposed method was compared with classical framework based algorithms RANSAC, PROSAC, MAGSAC++, GMS, AdaLAM, PEARL, MTC, Sequential RANSAC, and deep learning based algorithms SuperGlue, OANet, CLCNet, CONSAC, etc. Results indicate over 30% increase in the inlier rate, 8.39% reduction in the mis-classification rate of multi model estimation. It is concluded that the new algorithm has significant advantages in incorrect matches filtering and multi-model estimation. [ABSTRACT FROM AUTHOR]

Current researches on NDP flooding behavior detection mainly focus on detecting RA flooding and NS flooding behaviors, and there is insufficient flooding detection for other messages of the NDP protocol. Moreover, traditional threshold rule detection methods suffer from poor dynamics and low accuracy, while most of the Al-based detection methods can only perform binary classification detection, and there are still challenges in performing multi-classification detection. In short, there is a lack of corresponding research in multi-classification flooding detection of all messages of NDP protocol. Therefore, this paper proposed a multi-classification detection framework for NDP protocol flooding behaviors, and proposed a flooding behavior detection method for NDP protocol based on time interval characteristics. The framework constructed the first multiclassification dataset for NDP flooding detection through the processes of traffic collection and data processing, it compared and used 5 machine learning and 5 deep learning algorithms to train the detection model. The experimental results show that the detection accuracy of the XGBOOST algorithm in machine learning can reach 99.18%, and the detection accuracy of the Transformer algorithm in deep learning can reach 98.45%. Compared with the existing detection methods, the accuracy is higher. Meanwhile, the detection framework can detect 9 types of flooding behaviors for all 5 types of messages of NDP protocol and classify the flooding behaviors into multiple types. [ABSTRACT FROM AUTHOR]

For charging problem in AGV scheduling in automated container terminals, this paper constructed a mixed integer optimization model considering the shallow charging and shallow discharging charging strategy. The model aimed to minimize the final completion time of the AGV. Under the constraints of considering the change of AGV battery power and the difference in power consumption in different states of the AGV, the model used the AGV idle time and the end time of a work cycle to make up power, reducing the number of AGV charging times, and thus reducing the total completion time. The model was solved by Wolf-PHC reinforcement learning, and the results were compared with GAMS solver, Q-learning algorithm and genetic algorithm (GA) respectively to verify the effectiveness of the model and superiority of the algorithm. The example analysis shows that AGV utilization efficiency is higher under the shallow charging and shallow discharging charging strategy, and the combination of Wolf-PHC and GA is better for the model solution. [ABSTRACT FROM AUTHOR]

Threshold private set intersection (TPSI) is a special case of secure multi-party computation, which is widely used in many fields such as machine learning, ride-sharing, and fingerprint recognition. However, currently existing solutions are all based on algorithms with high computational complexity and are only implemented under a semi-honest model, resulting in high computational overhead and inability to resist attacks by malicious adversaries. In order to solve the above problems, this paper first proposed a vector oblivious match test(VOMT) protocol and designed an efficient semi-honest TPSI protocol based on VOMT and cuckoo hashing. In addition, it combined VOMT and symmetric key encryption schemes to construct the vector oblivious decryption matching test(VODMT) protocol and designed a TPSI protocol based on VODMT and oblivious pseudorandom functions that could resist malicious adversaries. Subsequently, it proved the security of the protocol under the semihonest model and the malicious model respectively, and analyzed that the computational complexity and communication complexity of the two protocols were linear. When the set size is 4 096, the online running time of the two proposed protocols is 0.81 s and 1.81 s respectively, while the previous work required 5 627 s, so both protocols are efficient. [ABSTRACT FROM AUTHOR]

Federated learning, a novel distributed machine learning paradigm, collaboratively trains a global model while preserving data privacy. It faces challenges of slow convergence and low accuracy in the global model under data heterogeneity. Aiming at the problem, this paper proposed a personalized federated multi-task learning optimization (FedMTO) algorithm tailored for heterogeneous data. In a multi-task learning framework that included global and local tasks, it considered the personalized federated optimization problem. Initially, FedMTO adopted the idea of parameter decomposition, coordinating global and local classifiers through the learning of adaptive classifier combination weights. This process extracted knowledge from global classifiers to achieve personalized modeling for local tasks. Furthermore, due to the varying data distributions of local tasks, FedMTO incorporated a regularization multi-task learning strategy during local updates. This approach focused on the relevance between tasks to reduce the differences among various local tasks, thus ensuring fairness in the federated learning process. Finally, experiments were conducted on the MNIST and CIFAR-10 datasets under different data heterogeneity scenarios. The results demonstrate that compared with existing algorithms, FedMTO achieves higher accuracy and better fairness, verifying the effectiveness of this method in addressing heterogeneous data problems in federated learning. [ABSTRACT FROM AUTHOR]

In order to solve the problem that class-imbalance data of clinical medical tables tend to have an impact on the model and that the performance of deep learning framework is difficult to match that of traditional machine learning methods when processing scale data tasks, this paper proposed a layer by layer Transformer (LLT) network model based on cascaded under-sampling. LLT deleted the most types of data layer by layer by cascade under-sampling method to achieve the balance of data categories and reduced the impact of class-imbalance data on the classifier. Moreover, LLT used attention mechanism to carry out correlation evaluation on the features of the input data to achieve feature selection, refined the feature extraction abi-lity and improved the model performance. This paper used RA (rheumatoid arthritis) data as test samples. Experimental results show that, on the premise of not changing the sample distribution, the recognition rate of a few categories is increased by 6.1% by the proposed cascade under-sampling method, which is 1.4% and 10.4% higher than that of the commonly used NEARMISS and ADASYN respectively. The accuracy of the RA tabular data and the F1-score index of LLT reach 72.6% and 71.5%, the AUC value is 0.89, the mAP value is 0.79, and the performance exceeds the current mainstream tabular data classification models such as RF, XGBoost and GBDT. This paper also visualized the model process and analyzed the characteristics affecting RA. It has a good guiding significance for the clinical diagnosis of RA. [ABSTRACT FROM AUTHOR]

As a classification task, the detection of new malicious users in social networks has been facing the lack of datasets and labels of malicious users. With limited data, this paper proposed a method based on adaptive differential graph convolution to detect malicious users accurately. By extracting user features and social relationships in the social network, the method constructed the social network graph. After this, it calculated the similarities between node and its neighbors to prioritize the neighbors, and used the priority order to sample key neighbors. The node used adaptive weighted average to aggregate the features of key neighbors to itself, to update its features. After feature updating, by feature dimension reduction and normalization, the node got its malicious value, for malicious detection. The experiment results show that, compared to other methods, the method proposed in the paper achieves higher precision and overall accuracy on detection of new malicious users, with a satisfactory speed. Results also demonstrate that adaptive differential graph convolutional networks can effectively capture the key features of a small number of data samples. [ABSTRACT FROM AUTHOR]

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.)

In federated learning environments, it is crucial to assess the quality of client data, especially when a validation set is not available. Traditional evaluation methods rely on measuring the loss of client models on the validation set of a central node to assess data quality. To address these issues, this paper proposed a method for scoring model quality based on peer information. This method involved tailoring the model parameters uploaded by the client and designing a model quality scoring mechanism based on the theories of correct scoring rules. It developed an optimized aggregation algorithm, leveraging the scores of clients to mitigate the impacts of low-quality local models on the global model. Experiments conducted on datasets like MNIST, Fashion-MNIST, and CIFAR-10 demonstrate that the proposed scoring mechanism is straightforward and effective in identifying three types of low-quality data clients: free-riding clients, overly privacy-protective clients, and mislabeled clients. The proposed method enhances the robustness of federated learning performance. [ABSTRACT FROM AUTHOR]

Federated learning solved the data silo problem of machine learning. However, the dataset of each party may have large differences in the instance space and feature space, which led to the degradation of prediction accuracy of the federated model. To address the above problems, this paper proposed a federated learning method based on differential privacy protection knowledge transfer. The method used boundary-expanding locality-sensitive hashing to calculate the similarity between instances of each party, and carried out weighted training of instances according to the similarity to achieve instance-based federated transfer learning. In the above process, each party needs not to disclose their instances to other parties, preventing the direct leakage of privacy. Meanwhile, to reduce the indirect privacy leakage in the knowledge transfer process, the proposed method introduced differential privacy mechanism to perturb the gradient data transmitted between all parties, so as to achieve privacy protection in the process of knowledge transfer. Theoretical analysis shows that the knowledge transfer process satisfies ε-differential privacy protection. This paper implements the proposed method based on the XGBoost model. The experimental results show that, compared with the other method without knowledge transfer, the proposed method reduces the test error of the federated model by more than 6% on average. [ABSTRACT FROM AUTHOR]

In the practical application of recommendation systems, item popularity bias can be amplified by feedback loops, machine learning training models, and some external factors. It results in a phenomenon where a large number of long-tail items do not get a fair chance to be recommended. To address the fairness problem caused by the feedback loop amplifying the popularity bias, this paper conducted the first fairness analysis and enhancement study by means of a stochastic model checking method. It modeled the traditional recommendation system framework based on popularity bias and feedback loops as a DTMC and verified the fairness properties. The experiment revealed that as the number of feedback loop rounds increased, the Matthew effect intensified and fairness significantly diminished. This paper presented a fairness-enhanced recommendation system framework (FERSF) guided by stochastic model checking. It added a dynamic fairness threshold detection process to the feedback loop of the traditional framework to monitor the fairness. Also, it made a fairness-enhanced adjustment of the feedback influence factor to mitigate the impact of popularity bias on the system. The experimental analysis shows that the fairness of FERSF is significantly improved compared to the traditional recommendation system. Compared with the methods based on utility functions for fairness improvement, FERSF fundamentally inhibits the amplification of popularity bias due to the dynamic nature of the combined feedback loop. Compared with other algorithm-specific fairness improvements, FERSF is highly compatible because it is modeled based on the recommendation system framework. [ABSTRACT FROM AUTHOR]

In recent years, various human action recognition algorithms have achieved excellent recognition performance based on a large number of labeled samples. However, in practical applications, acquiring training samples and their corresponding labels is time-consuming and laborious, which limits the actual implementation of the algorithm. This paper summarized the deep learning algorithms for action recognition under weak supervision and few-shot learning. Firstly, in the case of weak supervision, it classified and summarized the semi-supervised action recognition methods and unsupervised domain adaptation video action recognition methods separately. Then, it reviewed video action recognition algorithms based on few-shot learning in detail. Further, it summarized the relevant human behavior datasets and analyzed and compared the performance of various relevant video action recognition algorithms on these datasets. Finally, the paper summarized the full text and discusses the future development trend of human action recognition. [ABSTRACT FROM AUTHOR]

Continual learning, as a special machine learning paradigm that continuously learns new tasks in non-stationary data streams and can maintain the performance of old tasks, is a hot research topic in fields such as visual computing and autonomous robotics, but at this stage, the catastrophic forgetting problem is still a great challenge for continuous learning. This paper conducted a review study on the catastrophic forgetting problem of continual learning, analyzed the mechanism of catastrophic forgetting problem mitigation and explored the catastrophic forgetting problem solving strategies at three levels, included regularization strategy, replay strategy, dynamic architecture strategy and joint strategy, in terms of model parameters, training data and network architecture. According to the existing literature, this paper condensed the evaluation index of the catastrophic forgetting method and compared the performance of solving strategies for different catastrophic forgetting problems. Finally, it pointed out the future research direction of continual learning, to provide references for the study of continuous learning catastrophic forgetting problems. [ABSTRACT FROM AUTHOR]

Aiming at the hybrid flow shop scheduling problem with the objective of minimizing the makespan, this paper proposed an opposite artificial bee colony algorithm with opposition-based learning to solve the problem. First, according to the characteristics of hybrid flow shop scheduling problem, this paper established the corresponding mathematical model and simulation optimization model. Secondly, in order to avoid the problem of local optima in the process of optimization, this paper added the opposition-based learning strategy in the three stages of population initialization, employment bee and observation bee, used the reverse order strategy between two points and the element exchange strategy to speed up the optimization, then used the elitist preservation strategy to retain the optimal solution. Finally, the simulation experiment selected two examples and 21 benchmark problems of different scales, the results of simulation experiment verified that proposed algorithm is effective in solving such problems by comparison with the experimental of related algorithms. [ABSTRACT FROM AUTHOR]

In recent years, the deep learning network model based on UNet and GAN (generative adversarial network) has shown unique advantages in the field of image restoration. But, there are still artifacts, blurring, degradation of texture details, difficult to repair large-area damage, and the repaired holes are incompatible with the background image. In order to solve the problems that the current existing models are not friendly to the repair of large-area damaged images and the degraded images after repair, this paper improved the Shift-UNet model by studying the existing methods. On the basis of U-Net and GAN, this paper added an improved attention mechanism Attention-UNet between each layer of encoder and decoder and integrated it into the Shift-UNet to form Attention-Shift-UNet. And through research, it changed the activation function of the original down-sampling part from Leaky＿ ReLU to SiLU. The improved model not only achieved good results on the 64×64 central cover, but also realized the random cover, and the cover area increased from 20 % to 80 %. The experimental results show that the repair effect of the proposed model is better, especially for the repair of large-area damaged images. After testing on CelebA, Paris Architecture and Paris Stree View datasets, the evaluation indexes have been significantly improved, and SSIM has increased from 0.944 5 to 0.947 1. PSNR is increased from 27.992 7 to 28.553 6. L2 loss is decreased from 0.001 7 to 0.001 5. [ABSTRACT FROM AUTHOR]

Comparing the similarity of different enhancements of the same image is the key to achieving remarkable results in contrastive learning. The traditional contrastive learning method uses two different views of the image. In order to learn more information of the image to improve the classification accuracy, based on MoCo, this paper presented multi-view momentum contrast learning algorithms. In each iteration, this paper used a query encoder and multiple momentum encoders to extract feature for multiple data enhancements of the image, so that it could use more data enhancements and negative samples in this iteration. And it used the optimized noise contrast estimation(InfoNCE) to calculate the loss, so that the query encoder could obtain feature representations which was more beneficial to downstream tasks. The query encoder network updated with gradient backhaul, and each momentum encoder updated with an improved momentum update formula to improve the generalization ability of the model. Experimental results show that using multi-view momentum contrastive learning can effectively improve the classification accuracy of the model. [ABSTRACT FROM AUTHOR]

Aspect and opinion pair extraction aims to extract aspect and opinion items and match relations from a given sentence. However, related studies typically extract aspects and opinions independently without identifying the relationships. To identify the relationships of aspect and opinion item, this paper proposed a knowledge-augmented multi-task learning model for aspect and opinion pair extraction. First, it used the pre-trained language model to generate word vectors with semantic information for the text. In order to achieve the effect of knowledge enhancement, it used the masked attention mechanism to integrate the semantic information of the knowledge graph into the word vectors, and used the sequence labeling method based on the distance-based attention and conditional random fields to extract aspects and opinions. Finally, it matched the extracted aspects and opinions to predict the corresponding relationship. In order to strengthen the connection between the aspect and opinion extraction module and the matching module, the model adopted a shared coding layer to achieve joint training. In addition, in the training process, the matching module used the real labels as input, and used the result of the extraction module as input in the testing process. Finally, to demonstrate the effectiveness of the model, this paper used three general domain datasets for comparative experiments. The model achieves F, values of 66. 99%, 75. 17% and 67. 30% in aspect and opinion matching tasks respectively, and outperforms other comparative models. [ABSTRACT FROM AUTHOR]

In order to further mine the changelog information, this paper proposed a siamese Bi-lSTM network model based on the attention mechanism to classify the changelog to realize topic annotation and assist in locating the location of code defects. The model proposed a word segmentation tool with security features to realize changelog preprocessing, used Bi-lSTM network to learn changelog contextual semantic information, and solved the problem of overfilling mode existing in changelog itself through siamese neural network and expanded the data set with high quality to improve generalization ability. It carried out serialization training for the changelog composed of multiple sentences, and the influence of sentences was distinguished through the attention mechanism. For some changelogs of the defect repair class, this paper improved the LLVM tool, generated a mapping table to search the log content, and located the location of the defect module in the source code. A large number of experimental results show that the classification effect of the model in this paper has strong generalization ability, and is nearly 10% higher than the general methods in text classification methods in terms of accuracy, F, value and other indicators, has ideal log classification effect and topic learning effect. [ABSTRACT FROM AUTHOR]

Aiming at the deficiency that K2 algorithm depends on the maximum number of parent nodes and node of order, this paper proposed an improved Bayesian network structure learning algorithm (MWST-CS-K2). The algorithm firstly obtained the maximum number of parent nodes by constructing the maximum support tree (MWST). Then it used the correlation degree and updated coefficient between variables to set rules for adding, subtracting and reversing edges. The paper used improved cuckoo algorithm to optimize the nest location, and applied the breadth first search strategy to search and traverse to get the node order. Finally, the paper used maximum number of parent nodes and the order of nodes as the input of K2 algorithm to get the final network. Experiments show that in the standard ASIA,SACHS and CHILD network data tests, the average correct edge ratio of MWST-CS-K2 algorithm reaches 97.3%, 87.7 % and 95.6% respectively. The learning effect is better than other corresponding comparison algorithms, and the network structure obtains the most similar to the standard network structure. [ABSTRACT FROM AUTHOR]

To address the problem of poor recommendation accuracy in the session recommendation model, this paper proposed a capsule-based graph convolutional disentanglement for session-aware recommendation (CGCD) method.Specifically, it used disentangled learning technique to transition item embedding into factor embedding based on multiple sub-channels. Then, it used a capsule dynamic fusion strategy to aggregate different factors to obtain a new item embedding.In addition, it used a multi-head attention mechanism to assign weights to each item in the session. Finally, it aggregated the item embedding with other items in the current session according to the assigned weights, and then generated an accurate session representation to achieve item recommendation. Experiments on two publicly available real datasets show that the proposed model can improve the Pre@10, Pre@20, MRR@10 and MRR@20 by 5.17%, 2.99%, 6.56% and 2.94% on average of the recommended performance, which verifies the effectiveness and efficiency of the method in this paper. [ABSTRACT FROM AUTHOR]

Most of the existing image tampering detection methods are only for a certain tampering method, and there is a problem that the detection accuracy of the tampered area boundary is not high. In this regard, this paper proposed a U-shaped network based on dual-stream encoder-decoder architecture for image tampering. Firstly, the method used the skip connection between the encoder and the decoder to fuse the low-level and high-level features in the tampered image, and used the atrous convolution and CBAM attention mechanism to fuse the features output by the encoder, so that the network had better localization performance for tampered regions of different scales; Secondly, in order to improve the network’s detection accuracy of the boundary of the tampered area, this algorithm used the image morphological method to make a tampered boundary dataset; Finally, it used multiple loss functions to optimize the performance of the network simultaneously, namely it used cross-entropy and root-mean-square loss functions to measure the tampered region loss and tampered boundary loss of the prediction graph respectively. The experimental results on 4 public datasets: CASIA, Columbia, NIST16, Coverage show that the method proposed in this paper can effectively detect the tampering area of the fake image with splicing and copy-paste tampering methods, and output pixel-level tampering region location map. Compared with other mainstream tampering detection methods, the proposed method achieves the highest AUC values on the CASIA and Columbia dataset and the highest F1 values on the Columbia dataset. [ABSTRACT FROM AUTHOR]

In order to solve the problem of low accuracy of 6D pose estimation for textured models under occlusion, this paper proposes an end-to-end 6D pose estimation algorithm based on local feature representation. Firstly, this paper proposes a Spatial and Coordinate Attention mechanism to obtain accurate localization information. A YOLOv5-CBE detection network forms by adding the attention mechanism to the backbone network and introducing a weighted Bidirectional Feature Pyramid Network in the detection layer. The Precision, Recall and mAP@0.5 of YOLOV5-CBE algorithm rise by 3.6%, 2.8% and 2.5% respectively, and the coordinate error of local feature center point decreases by 25% at most. Secondly, the YOLOv5-CBE network detects the local feature key points and calculates 6D pose of the model with 3 D Harris key points by Singular Value Decomposition, and the algorithm can guarantee 2D reprojection Accuracy and ADD Accuracy above 95% with 70% occlusion, which has a strong robustness. [ABSTRACT FROM AUTHOR]

Network fault management aims to detect, identify and correct error conditions occurring in the network, providing users with reliable and stable network services. Recently, the utilization of machine learning methods for network fault management has attracted widespread attention. In this regard, this paper first introduced the research background of cell network fault management and explain the process and function of network fault management, and then introduced the existing cell network fault management frameworks. Subsequently, this paper reviewed the existing research on machine learning methods in cell network fault management. Referring to the fault management lifecycle, this paper introduced, summarized and compared the machine learning methods for fault detection, fault diagnosis and fault prediction for the purpose of providing reference for research in related fields. [ABSTRACT FROM AUTHOR]

In reality, the data generated in many fields usually have multiple classes and are imbalanced. In multi-class imbalance classification, problems such as class overlap, noise and multiple minority classes reduce the capability of classifiers, and effective solution of multi-class imbalance problem has become an important research topic in the field of machine learning and data mining. Based on the recent literature on multi-class imbalance classification methods, this paper analyzed and summarized both data preprocessing and algorithm-level classification methods, and conducted a detailed analysis of all algorithms in terms of advantages, disadvantages and data sets. The data preprocessing methods introduced oversampling, under-sampling, hybrid sampling and feature selection methods to compare the performance of the algorithms using the same datasets. In addition, the algorithm-level classification methods described and analyzed base classifier optimization, ensemble learning and multi-class decomposition techniques. Finally, this paper summarizes the future development directions of the multi-class imbalanced data classification research field. [ABSTRACT FROM AUTHOR]

Aiming at the problems of low detection accuracy in massive unbalanced data and poor model generalization ability caused by data heterogeneity in traditional anomaly detection methods in IoT scenarios, this paper proposed an adversarial dual-coding anomaly detection network based on federated learning(GB-AEnet-FL). Firstly, this paper proposed an active feature distribution learning algorithm based on abnormal data, which actively learns the potential feature distribution of data, expands abnormal data through data reconstruction, and balances the proportion of positive and negative samples. Secondly, this paper introduced an adversarial training mechanism in the latent feature layer adding consistency enhancement constraints and shrinkage constraints to improve the accuracy of feature extraction. Finally, this paper designed a federated learning algorithm based on dynamic model selection which compares the confidence scores of the local model and the global model, dynamically selects part of the federated bodies to participate, accelerates the aggregation of models, and protects user privacy to a certain extent. Validated on four different datasets, the results show that the proposed algorithm is better than the traditional algorithm in detection accuracy, and the generalization ability is improved accordingly. [ABSTRACT FROM AUTHOR]

Multiple random empirical kernel learning machine(MREKLM) selects a small number of samples to construct the empirical feature space, but does not take into account the distribution of data when projecting, and takes a long time to select samples. In order to utilize the distribution information of samples, this paper introduced the intraclass dispersion matrix, and proposed an intra-class variance based multiple empirical kernel learning(ICVMEKL), which enabled samples to take into account the intra-class information when projecting. The method enhanced the classification boundary and improved the classification accuracy. Further, in order to reduce the selection time of samples, this paper proposed an improved ICVMEKL(ICVMEKL＿I) based on locality-sensitive hashing algorithm by using the border point extraction based on locality-sensitive hashing(BPLSH) to select samples, so that the samples for constructing the empirical kernel no longer need to be obtained from the candidate set, and the method reduced the training time. Experiments on multiple datasets show that ICVMEKL can effectively improve the accuracy, ICVMEKL＿I can significantly reduce the training time, both of which show good generalization performance. [ABSTRACT FROM AUTHOR]

Complementary product recommendation is to recommend products to users that can be used together. Existing product recommendation methods consider all features of product’s images and texts, but do not consider the relationship between visual and textual modalities, and not all features contribute equally to the complementary relationship. Based on this situation, this paper proposes a complementary product recommendation model fusing with attention mechanism, dubbed as CPRFA. The model first enriches its feature representation with the product’s image and text information, and secondly, in order to unify heterogeneous products with multimodal information, it further converts them into image and text vector representations, and then uses the attention mechanism to adaptively assign weights to different features of the product, and finally learns the relationship between the product and the candidate products for the recommendation of complementary products. In the experiments based on the Amazon dataset, the CPRFA model proposed in this paper outperforms other baseline models in all indicators. Experiments show that the CPRFA model can recommend complementary products to users more accurately. [ABSTRACT FROM AUTHOR]

The U-turn task is one of the contents of autonomous driving research, and most of the solutions under the standard roads in cities cannot be implemented on non-standard roads. Aiming at solving this problem, this paper establishes a vehicle U-turn dynamical model and designs a multi-scale convolutional neural network to extract feature maps as the input of the agent. In addition, for the sparse reward problem in the U-turn task, this paper proposes a hierarchical proximal policy optimization algorithm that combines hierarchical reinforcement learning and proximal policy optimization algorithm. In experiments with simple and complex scenarios, this algorithm learns policies faster and has a higher success rate of U-turn compared to other algorithms. [ABSTRACT FROM AUTHOR]

In order to overcome the challenges of policy learning in Multi-Agents Systems（MASs）, such as the unstable environment and the interaction of agent decisions, this paper proposes a method named Observation Relation Extraction（ORE）, which uses a complete graph to model the relationship between different parts of each agent’s observation, and takes advantage of the attention mechanism to calculate the importance of the relationship between different parts of each agent’s observation. By applying the above method to multi-agent reinforcement learning algorithms based on value decomposition, this paper proposes multi-agent reinforcement learning algorithms based on observation relation extraction. Experimental results on StarCraft Multi-Agent Challenge（SMAC） show the proposed algorithms with ORE leads to better performance than the original algorithms in terms of both convergence speed and final performance [ABSTRACT FROM AUTHOR]

Facing the current situation of load uncertainty, new energy grid integration, and “dual carbon” target in the power system, the paper established a real-time pricing model considering load uncertainty and carbon trading in the context of the smart grid with full consideration of the welfare of both supply side and user side. Based on the advantages that reinforcement learning can handle variable complexity, non-convex, and nonlinear problems, this paper used the Q-learning algorithm in reinforcement learning to solve the model iteratively. Firstly, this paper transformed the real-time interaction process between the user and the power supplier into a Markov decision process corresponding to the reinforcement learning framework. Secondly, the process represented the information interaction between the user and the power supplier as the iterative exploration of the agent in a dynamic environment. Finally, this paper found the optimal value by the Q-learning algorithm in reinforcement learning, i. e., the maximal social welfare value. The simulation results show that the proposed real-time pricing strategy can effectively enhance social welfare and reduce total carbon emissions, which verifies the feasibility and effectiveness of the proposed model and algorithm. [ABSTRACT FROM AUTHOR]

With the emergence of various delay-sensitive applications, how to improve the system's delay performance has be come a hot topic in academic. However, it is difficult for existing multi-domain embedding algorithms to meet the delay per- formance requirements of virtual networks. Therefore, to solve this problem, this paper proposed a TDS-VNE algorithm. In the node embedding phase, it defined a propagation delay evaluation function(D). In the link embedding phase, it defined a path delay sensing parameter. Simulation results show that the proposed embedding algorithm reduces the average network propaga tion delay and has good results on the virtual network request acceptance rate and long-term benefit-cost ratio [ABSTRACT FROM AUTHOR]

The compressive strength of concrete is a crucial parameter in the design and evaluation of building structures,as it directly impacts the quality and safety of buildings. To solve the current problem that the prediction of concrete compressive strength by the existing machine learning models suffers from issues such as long prediction times and low precision, which cannot meet the real-time and accuracy requirements of the prediction on construction sites, this paper proposes a novel method for predicting concrete compressive strength based on the Tuna Swarm Optimization Algorithm（TSO-ELM）. The proposed method utilizes TSO to optimize the offset value of the connection weight in the initial parameters of the ELM hidden layer, thereby enhancing the prediction accuracy of ELM. In the simulation experiment, the prediction speed of ELM, the optimization ability of TSO, and the generalization of the TSO-ELM model are verified using two sets of concrete data. The results demonstrate that the proposed method can significantly improve prediction speed and accuracy with fewer iterations and better generalization, providing a new approach for the timely prediction of concrete compressive strength in field construction. [ABSTRACT FROM AUTHOR]

Local SGD is used in distributed machine learning to alleviate communication bottleneck. However, it increases the time gap between heterogeneous cluster workers in local computation with the characteristics of local multi-round iteration, which bring about large synchronization wait time and the problem of stale parameters. Regarding the above problems, based on the Local SGD algorithm, this paper proposed a dynamic partial synchronization communication method（LPSP）. This method makes full use of the advantage of Local SGD with two-layer decision. After each worker finishes each epoch of local gradient calculation, it determines its communication possibility based on local training situation. And parameter server divides the synchronization worker sets to minimize the synchronization wait time. Based on above strategies, the communication overhead of Local SGD method is reduced and the negative influence of straggler is effectively controlled. Experiments show that LPSP can achieve up to 0.75 to 1.26 times acceleration without loss of training accuracy, and up to 5.14% accuracy improvement, which can effectively speed up the convergence of model training. [ABSTRACT FROM AUTHOR]

Aiming at the problems of poor generalization performance and overlapping between classes of existing cross scene face anti-spoofing detection models, this paper proposed a face anti-spoofing method based on conditional adversarial domain generalization. Firstly, the method used U-Net network embedded with attention mechanism and ResNet-18 encoder to extract features from multiple source domains, then sent the extracted features to auxiliary classifier, and combined the output of the feature encoder with the prediction results of the classifier by the method of multilinear mapping, and then input into the domain discriminator for adversarial training to achieve feature and class-level alignment of multiple source domains. Secondly, in order to reduce the impact of difficult to transfer samples with inaccurate prediction on domain generalization, this paper used entropy function to control the priority of samples to improve the performance of domain generalization. In addition, by adding a face depth map to further capture the distinguishing features of the real and attack face, and using asymmetric triplet loss constraint as auxiliary supervision to further improve the intra-class compactness and inter-class discrimination. Comparative experiments on public live detection datasets verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Since edge clouds are not more powerful than the central cloud, it is easy to lead to unexpected autoscaling or low resource processing capabilities in response to dynamic workload. Therefore, this paper used the microservice application in a real edge cloud environment to experimentally evaluate two synthesis and two actual workloads. And this paper proposed a hybrid autoscaling method based on workload prediction（Predictively Horizontal and Vertical Pod Auto-scaling, Pre-HVPA）, which first uses machine learning to carry out the workload data characteristics. After obtaining the final workload prediction result, this paper use the predictive workload for hybrid autoscaling module. The simulation shows that the microservice autoscaling policy based on this method can reduce more scaled jitter and more pod container use, and the method is scalable, so it is suitable for the microservice applications in edge cloud environment. [ABSTRACT FROM AUTHOR]

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.)

In dynamic gesture recognition methods based on deep learning, aiming at the problems of large network scale, simple dimension of extracted spatiotemporal features, and insufficient extraction ability of effective features, this paper proposed a deep network framework. Firstly, this paper applied a novel spatiotemporal convolution module based on multiscale information fusion in the time domain to improve the 3D residual network structure, greatly reduced the size of the network and obtained rich spatiotemporal receptive field characteristics. Then it introduced a spatiotemporal feature channel attention mechanism with global information synchronization, and used a few parameters to construct the global dependency between feature maps, and the module could obtain the key features of dynamic gestures more efficiently. The experimental results on the self-built gesture dataset SHC and the public gesture dataset SKIG show that the proposed gesture recognition method has fewer parameters and more powerful multiscale spatiotemporal feature extraction ability, and gains a higher recognition rate than the current mainstream algorithms. [ABSTRACT FROM AUTHOR]

Federated learning is a novel distributed machine learning approach, which provides a privacy protection way to learn a shared model without sharing each client's private data. However, the existing frameworks of federated learning only work for supervised learning wherein each client's data is labeled. Since collecting labeled data is difficult and expensive to obtain in the real world, the assumption of federated learning is not valid. To solve this problem, this paper proposed a semi-supervised federated learning model named ANN-SSFL based on an AutoEncoder neural network. The proposed model was extended from classical federated learning and allowed clients who might not have labeled data to participate the federated learning. The latent features which could be identified by the classifier were obtained by AutoEncoder neural network from unlabeled data, therefore unlabeled data could provide their data information to make their contributions. This paper conducted experiments on MNIST data sets. The experimental results show that the proposed ANN-SSFL is practical and effective. When the number of supervised clients remains unchanged, adding unsupervised clients can improve the accuracy of classical federated learning. [ABSTRACT FROM AUTHOR]

Facial local detail information plays an important role in facial expression recognition（FER）.However, most of the existing methods only focus on the high-level semantic information of facial expressions, while ignoring the fine-grained information of local facial regions. To solve this problem, this paper proposed a deep convolutional neural network fusing local feature and two-stage attention weight learning（FLF-TAWL）,which could adaptively capture important facial regions to improve the effectiveness of facial expression recognition. The FLF-TAWL model was composed of a dual-branch framework, one branch extracted local features from image blocks, and the other branch extracted global features from the entire expression image. Firstly, this paper proposed a two-stage attention weight learning strategy. In the first stage, it roughly learned the importance weights of global and local features, in the second stage, it further refined the attention weight, and fused the local and global features. Secondly, the model used a region-biased loss function to encourage the most important regions to obtain higher attention weights. Finally, this paper carried out extensive experiments on FERPlus, Cohn-Kanada（CK+） and JAFFE datasets to obtain accuracy rates of 90.92%,98.90% and 97.39% respectively. The experimental results verify the effectiveness and feasibility of the FLF-TAWL model. [ABSTRACT FROM AUTHOR]

In aspect-level sentiment analysis research, previous work often ignores the problem that the contribution of different types of parts of speech and the interaction of local and global features affects the classification accuracy. This paper proposed an aspect-level sentiment analysis model DPMHA based on double-layer part-of-speech-aware and multi-head interactive attention mechanism. Firstly, this paper used BERT pre-training model to obtain word vectors of contextual information. Secondly, this paper creatively proposed two part-of-speech-aware local feature extraction layers, which could focus on these words around aspect words with important parts of speech and reduce the influence of noise words. Then, this paper designed a multi-head interactive attention mechanism between local features and global features to fully explore the important interactive features between them. Finally, this paper also proposed a dynamic feature fusion layer and softmax layer to obtain the results of sentiment analysis. Experiments on three public data set show that compared with the existing aspect-level sentiment analysis model, the DPMHA model increases the MF1 value of the restaurant14,laptop14,and restaurant15 datasets by 2.41%,1.24% and 2.39% respectively, and the accuracy rate increases by 1.34%,0.78% and 0.37% respectively. [ABSTRACT FROM AUTHOR]

Due to the lack of sufficient co-occurrence information in short text, weak connections between words, and it is difficult to obtain subject words, which leads to the need to manually label a large number of training samples for short text classification, and the problems of sparse features and dimension explosion. This paper proposed a word symbiotic short text classification model based on attention mechanism and label graph（WGA-BERT）.Firstly, this paper used the pretrained BERT model to calculate the context aware text representation, and used WNTM to model the potential word group distribution of each word to obtain the topic expansion feature vector. Secondly, this paper used a tag graph construction method to capture the structure and relevance of subject words. Finally, this paper used an attention mechanism to establish the relationship between subject words and between subject words and text, which solved the problems of data sparsity and subject text heterogeneity. The experimental results show that the WGA-BERT model improves the classification accuracy by an average of 3% compared with the traditional machine learning model. [ABSTRACT FROM AUTHOR]

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.)

Aiming at the problems of the current semantic segmentation network's spatial and channel feature mismatch, as well as the pixel loss of small target objects, this paper designed a dual-path semantic segmentation algorithm based on spatial feature extraction and attention mechanism. The spatial information path used four times downsampling to retain high-resolution features, and introduced a spatial feature extraction module to fuse multi-scale spatial information, thereby strengthening the network's ability to recognize small target objects. In addition, it used a semantic context path combined with two-stage channel attention to extract discriminative features, so that deep features could guide shallow features to capture more accurate semantic information, thereby reducing accuracy loss. This paper verified the algorithm on the CamVid dataset and Aeroscapes dataset, the mean intersection over union can reach 70. 5% and 51. 8% respectively. Compared with the current mainstream dual-path semantic segmentation model, the results verify the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

To address the problems of inefficient operation,excessive computational parameters, and high computational complexity in the Transformer-based machine translation model, this paper proposed a separate structure machine translation model based on fused CNN and Transformer. Firstly, for the problems of inefficient operation and excessive computational parameters, this paper used the structure of separating computational attention module and normalization module to ensure the reusability of stacked multilayer structure, improve the operation efficiency and reduce the computational parameters. Secondly, the model introduced the convolutional computation module and the original self-attentive module for fusion. This paper used the original self-attentive module to calculate the global contextual semantic relations and the convolutional computation module to calculate the local contextual semantic relations to reduce the complexity of the model. Experimental comparisons with other machine translation models on the same dataset show that the model has the lowest computational parameters and performs better than other models. [ABSTRACT FROM AUTHOR]

Poor text reading ability and inadequate visual reasoning were the main reasons for the insufficient effect of existing visual question answering models. To solve the above problems, this paper designed a MRGNN model. It used various forms of information in images to help understanding the scene text content, preprocessed the scene text image into the visual object graph and text graph respectively, and filtered the redundant information in the question self-attention module. It used an aggregator with attention to perfect the node features between subgraphs and fuse different modality information. The updated nodes used the context information of different modules to provide a better function for answering module. This paper verified the validity of MRGNN model on ST-VQA and Text VQA datasets. The experimental results show that MRGNN model achieves good results compared with some classical models for this task. [ABSTRACT FROM AUTHOR]

This paper applied the classical ID3 decision tree algorithm for discrete data classification. However, it required data discretization for continuous data processing, and this process easily caused information loss. This paper proposed the neighborhood equivalence relationship to induce the neighborhood ID3 ( NID3) decision tree algorithm. The NID3 algorithm improved the ID3 decision tree algorithm, which could directly implement continuous prediction and obtained better classification results. In the neighborhood decision system, this paper mined a neighborhood equivalence relationship. Then, based on the equivalent granulation of the neighborhood, this paper constructed the neighborhood information metric. Finally, this paper designed the NID3 decision tree algorithm based on the neighborhood information gain. As verified by both case analyses and data experiments, algorithm NID3 is effective for continuous data classification, and it outperforms ID3 algorithm in classification machine learning. [ABSTRACT FROM AUTHOR]

At present, researches on short-term load forecasting mainly focus on the analysis of influencing factors and the optimization of the model, and there are few researches on the robustness of the model. This paper took the extreme learning machine as the research object, conducted in-depth research on the robustness of the model, and applied it to the short-term load forecasting problem. The robustness of the ELM model was affected by the loss function. The current ELM model had poor robustness and low prediction accuracy when dealing with the samples containing outliers. To solve this problem, this paper proposed a new loss function based on p-order maximum correntropy criterion. It applied the loss function to the ELM model to improve its robustness in regression prediction. It proposed a method to estimate the percentage of noise in the actual sample, and gave the degree of noise pollution in the sample before established the load forecasting model. The simulation results show that the proposed algorithm model has better robustness and prediction accuracy when the outlier is more than 12%. [ABSTRACT FROM AUTHOR]

To meeting the exploding demand for computational resources in vehicles, offloading computational tasks to edge servers is allowed in vehicular edge computing. But how to make offloading decision and computational resource allocation are still critical issues that need to be addressed. Moreover, task unloading of moving vehicles in continuous time is rarely mentioned, especially the randomness of vehicle task arrival is not considered enough. To address the above problem, a dynamic vehicle edge computing model is established in this paper and is described as a Markov decision process in seven state two action spaces. A distributed deep reinforcement learning model is also built to solve the problem. Furthermore, for the discrete-continuous hybrid decision problem causing poor results, a deep reinforcement learning algorithm for split-order decision making is proposed, which nests the input layer with the first-order decision network. Simulation results show that the proposed algorithm has significant advantages in terms of task completion rate, time delay, and reward compared to the comparison algorithm by maintaining a lower level of energy consumption. This provides an effective solution to the offloading decision and computational resource allocation problem in vehicle edge computing. [ABSTRACT FROM AUTHOR]