检索结果-南通市图书馆

LMR-CBT: learning modality-fused representations with CB-Transformer for multimodal emotion recognition from unaligned multimodal sequences

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Computer science 2024年第4期18卷 39-47页

作者： Ziwang FU Feng LIU Qing XU Xiangling FU Jiayin QI school of Computer science(National Pilot Software Engineering school) Beijing University of Posts and TelecommunicationsBeijing 100876China Key Laboratory of Trustworthy Distributed Computing and Service(BUPT) Beijing 100876China Shanghai International school of Chief technology Officer East China Normal UniversityShanghai 200062China school of Computer science and technology East China Normal UniversityShanghai 200062China school of cyberspace Security Guangzhou UniversityGuangdong 510006China

Learning modality-fused representations and processing unaligned multimodal sequences are meaningful and challenging in multimodal emotion *** approaches use directional pairwise attention or a message hub to fuse language,visual,and audio ***,these fusion methods are often quadratic in complexity with respect to the modal sequence length,bring redundant information and are not *** this paper,we propose an efficient neural network to learn modality-fused representations with CB-Transformer(LMR-CBT)for multimodal emotion recognition from unaligned multi-modal ***,we first perform feature extraction for the three modalities respectively to obtain the local structure of the ***,we design an innovative asymmetric transformer with cross-modal blocks(CB-Transformer)that enables complementary learning of different modalities,mainly divided into local temporal learning,cross-modal feature fusion and global self-attention *** addition,we splice the fused features with the original features to classify the emotions of the ***,we conduct word-aligned and unaligned experiments on three challenging datasets,IEMOCAP,CMU-MOSI,and *** experimental results show the superiority and efficiency of our proposed method in both *** with the mainstream methods,our approach reaches the state-of-the-art with a minimum number of parameters.

关键词： modality-fused representations cross-model blocks multimodal emotion recognition unaligned multimodal sequences computational affection

Enhanced Differentiable Architecture Search Based on Asymptotic Regularization

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第2期78卷 1547-1568页

作者： Cong Jin Jinjie Huang Yuanjian Chen Yuqing Gong school of Computer science and technology Harbin University of Science and TechnologyHarbin150006China school of Automation Harbin University of Science and TechnologyHarbin150006China

In differentiable search architecture search methods,a more efficient search space design can significantly improve the performance of the searched architecture,thus requiring people to carefully define the search space with different complexity according to various *** rationalizing the search strategies to explore the well-defined search space will further improve the speed and efficiency of architecture *** this in mind,we propose a faster and more efficient differentiable architecture search method,***,we introduce a more efficient search space enriched by the introduction of two redefined convolution ***,we utilize a more efficient architectural parameter regularization method,mitigating the overfitting problem during the search process and reducing the error brought about by gradient ***,we introduce a natural exponential cosine annealing method to make the learning rate of the neural network training process more suitable for the search ***,group convolution and data augmentation are employed to reduce the computational ***,through extensive experiments on several public datasets,we demonstrate that our method can more swiftly search for better-performing neural network architectures in a more efficient search space,thus validating the effectiveness of our approach.

关键词： Differentiable architecture search allegro search space asymptotic regularization natural exponential cosine annealing

A Dual Discriminator Method for Generalized Zero-Shot Learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第4期79卷 1599-1612页

作者： Tianshu Wei Jinjie Huang school of Computer science and technology Harbin University of Science and TechnologyHarbin150006China school of Automation Harbin University of Science and TechnologyHarbin150006China

Zero-shot learning enables the recognition of new class samples by migrating models learned from semanticfeatures and existing sample features to things that have never been seen before. The problems of consistencyof different types of features and domain shift problems are two of the critical issues in zero-shot learning. Toaddress both of these issues, this paper proposes a new modeling structure. The traditional approach mappedsemantic features and visual features into the same feature space;based on this, a dual discriminator approachis used in the proposed model. This dual discriminator approach can further enhance the consistency betweensemantic and visual features. At the same time, this approach can also align unseen class semantic features andtraining set samples, providing a portion of information about the unseen classes. In addition, a new feature fusionmethod is proposed in the model. This method is equivalent to adding perturbation to the seen class features,which can reduce the degree to which the classification results in the model are biased towards the seen *** the same time, this feature fusion method can provide part of the information of the unseen classes, improvingits classification accuracy in generalized zero-shot learning and reducing domain bias. The proposed method isvalidated and compared with othermethods on four datasets, and fromthe experimental results, it can be seen thatthe method proposed in this paper achieves promising results.

关键词： Generalized zero-shot learning modality consistent discriminator domain shift problem feature fusion

Dynamic recrystallization mechanism, texture evolution development and mechanical characteristics of a Mg–8.7Gd–4.18Y–0.42Zr magnesium alloy by ECAP

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Progress in Natural science:Materials International 2024年第2期34卷 376-388页

作者： Ling Zhang Yinglong Li school of Materials science and Engineering Northeastern University school of Mechanical and Engineering Ningxia Institute of Science and Technology

Equal channel angle pressing is recognized for its ability to refine alloy grains and alter grain orientation, thereby achieve better mechanical performance of the magnesium alloy. This study investigates the microstructures,dynamic recrystallization mechanism, texture development, and mechanical performance of GW94K(Mg–8.7Gd–4.18Y–0.42Zr wt. %) Mg alloy following ECAP-4 passes at 400°C and 3 mm/min. Results show that when high-temperature deformation is undertaken, twin formation is suppressed while dislocation slip is facilitated, increasing dislocation density during deformation. Following ECAP deformation, the sample displayed higher fracture elongation, TYS, and UTS than the as-solutioned GW94K alloy. In particular, the GW94K alloy performed well mechanically after ECAP-4 passes, with an ultimate TYS of 231 MPa, an UTS of 290 MPa, and an elongation of 14.8 %. DDRX and shear bands induce CDRX, both of which are important in plastic deformation. as well as in modifying microstructure and grain orientation during ECAP deformation.

关键词： Dynamic recrystallization mechanism Equal channel angular pressing Magnesium alloy Mechanical performance Mg-Gd-Y-Zr alloy Texture development

Differentially Private Support Vector Machines with Knowledge Aggregation

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第3期78卷 3891-3907页

作者： Teng Wang Yao Zhang Jiangguo Liang Shuai Wang Shuanggen Liu school of cyberspace Security Xi’an University of Posts and TelecommunicationsXi’an710121China

With the widespread data collection and processing,privacy-preserving machine learning has become increasingly important in addressing privacy risks related to *** vector machine(SVM)is one of the most elementary learning models of machine *** issues surrounding SVM classifier training have attracted increasing *** this paper,we investigate Differential Privacy-compliant Federated Machine Learning with Dimensionality Reduction,called FedDPDR-DPML,which greatly improves data utility while providing strong privacy *** in distributed learning scenarios,multiple participants usually hold unbalanced or small amounts of ***,FedDPDR-DPML enables multiple participants to collaboratively learn a global model based on weighted model averaging and knowledge aggregation and then the server distributes the global model to each participant to improve local data *** at high-dimensional data,we adopt differential privacy in both the principal component analysis(PCA)-based dimensionality reduction phase and SVM classifiers training phase,which improves model accuracy while achieving strict differential privacy ***,we train Differential privacy(DP)-compliant SVM classifiers by adding noise to the objective function itself,thus leading to better data *** experiments on three high-dimensional datasets demonstrate that FedDPDR-DPML can achieve high accuracy while ensuring strong privacy protection.

关键词： Differential privacy support vector machine knowledge aggregation data utility

Dynamical localization in a non-Hermitian Floquet synthetic system

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2024年第5期33卷 147-151页

作者：可汗张嘉明霍良赵文垒 school of science Jiangxi University of Science and TechnologyGanzhou 341000China school of Information Engineering Jiangxi University of Science and TechnologyGanzhou 341000China

We investigate the non-Hermitian effects on quantum diffusion in a kicked rotor model where the complex kicking potential is quasi-periodically modulated in the time *** synthetic space with arbitrary dimension can be created by incorporating incommensurate frequencies in the quasi-periodical *** the Hermitian case,strong kicking induces the chaotic diffusion in the four-dimension momentum space characterized by linear growth of mean *** find that the quantum coherence in deep non-Hermitian regime can effectively suppress the chaotic diffusion and hence result in the emergence of dynamical ***,the extent of dynamical localization is dramatically enhanced by increasing the non-Hermitian ***,the quasi-energies become complex when the non-Hermitian parameter exceeds a certain threshold *** quantum state will finally evolve to a quasi-eigenstate for which the imaginary part of its quasi-energy is large *** exponential localization length decreases with the increase of the non-Hermitian parameter,unveiling the underlying mechanism of the enhancement of the dynamical localization by nonHermiticity.

关键词： Floquet system non-Hermitian physics dynamical localization

Federated Meta Reinforcement Learning for Personalized Tasks

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Tsinghua science and technology 2024年第3期29卷 911-926页

作者： Wentao Liu Xiaolong Xu Jintao Wu Jielin Jiang school of Computer science Nanjing University of Information Science and TechnologyNanjing 210044China school of Software Nanjing University of Information Science and TechnologyNanjing 210044China

As an emerging privacy-preservation machine learning framework,Federated Learning(FL)facilitates different clients to train a shared model collaboratively through exchanging and aggregating model parameters while raw data are kept local and *** this learning framework is applied to Deep Reinforcement Learning(DRL),the resultant Federated Reinforcement Learning(FRL)can circumvent the heavy data sampling required in conventional DRL and benefit from diversified training data,besides privacy preservation offered by *** FRL implementations presuppose that clients have compatible tasks which a single global model can *** practice,however,clients usually have incompatible(different but still similar)personalized tasks,which we called task *** may severely hinder the implementation of FRL for practical *** this paper,we propose a Federated Meta Reinforcement Learning(FMRL)framework by integrating Model-Agnostic Meta-Learning(MAML)and ***,we innovatively utilize Proximal Policy Optimization(PPO)to fulfil multi-step local training with a single round of ***,considering the sensitivity of learning rate selection in FRL,we reconstruct the aggregation optimizer with the Federated version of Adam(Fed-Adam)on the server *** experiments demonstrate that,in different environments,FMRL outperforms other FL methods with high training efficiency brought by Fed-Adam.

关键词： federated learning reinforcement learning meta-learning personalization

Dynamics of information diffusion and disease transmission in time-varying multiplex networks with asymmetric activity levels

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2024年第3期33卷 690-699页

作者：谢笑笑霍良安董雅芳程英英 Business school University of Shanghai for Science and TechnologyShanghai 200093China school of Intelligent Emergency Management University of Shanghai for Science and TechnologyShanghai 200093China school of Management Henan University of Science and TechnologyLuoyang 471023China

While the interaction between information and disease in static networks has been extensively investigated,many studies have ignored the characteristics of network *** this study,we construct a new two-layer coupling model to explore the interactions between information and *** upper layer describes the diffusion of disease-related information,and the lower layer represents the disease *** then use power-law distributions to examine the influence of asymmetric activity levels on dynamic propagation,revealing a mapping relationship characterizing the interconnected propagation of information and diseases among partial nodes within the ***,we derive the disease outbreak threshold by using the microscopic Markov-chain approach(MMCA).Finally,we perform extensive Monte Carlo(MC)numerical simulations to verify the accuracy of our theoretical *** findings indicate that the activity levels of individuals in the disease transmission layer have a more significant influence on disease transmission compared with the individual activity levels in the information diffusion ***,reducing the damping factor can delay disease outbreaks and suppress disease transmission,while improving individual quarantine measures can contribute positively to disease *** study provides valuable insights into policymakers for developing outbreak prevention and control strategies.

关键词： information diffusion disease transmission asymmetric activity levels quarantine strength

Protein crystallization with DNA templates

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Particuology 2024年第3期86卷 262-268页

作者： Bo Zhang Jia Yu Zhe Pu Wenhui Guo Ruoqi Yang Yu Wen Zhijiang Rong Hui Wang Xiaoli Wu Bo Wang Hongyan Sun school of Environment and Resources Taiyuan University of Science and TechnologyTaiyuan030024China school of Chemical Engineering and technology Taiyuan University of Science and TechnologyTaiyuan030024China

Protein crystallization plays a significant role in three-dimensional structural analysis and protein *** is important to increase the crystallization efficiency,which is possible by adding hetero-geneous templates in crystallization *** is biologically compatible and artificially designable polymer,which is easy to *** this study,single-and double-stranded DNA of precise sequences were designed and used as templates to promote protein crystallization of lysozyme and ***-fluence of DNA,single-stranded DNA with 10,20,40 bases and double-stranded DNA with 10,20,40 base pairs,were *** success rate of obtaining crystals of lysozyme and catalase in equal period was significantly improved with the addition of DNA comparing without templates ***-stranded DNA led to higher nucleation rate than that with single-stranded *** promotion of nucleation was more obvious at low concentration of protein solution and with longer chain DNA *** number and crystallization rate was enhanced with addition of long double-stranded DNA *** the results confirm that DNA is an effective polymer additive to enhance protein crystallization,especially for the application of the scarce protein crystallization.

关键词： DNA Polymer Additive Protein Crystallization