With the rapid expansion of social media,analyzing emotions and their causes in texts has gained significant importance.Emotion-cause pair extraction enables the identification of causal relationships between emotions...With the rapid expansion of social media,analyzing emotions and their causes in texts has gained significant importance.Emotion-cause pair extraction enables the identification of causal relationships between emotions and their triggers within a text,facilitating a deeper understanding of expressed sentiments and their underlying reasons.This comprehension is crucial for making informed strategic decisions in various business and societal contexts.However,recent research approaches employing multi-task learning frameworks for modeling often face challenges such as the inability to simultaneouslymodel extracted features and their interactions,or inconsistencies in label prediction between emotion-cause pair extraction and independent assistant tasks like emotion and cause extraction.To address these issues,this study proposes an emotion-cause pair extraction methodology that incorporates joint feature encoding and task alignment mechanisms.The model consists of two primary components:First,joint feature encoding simultaneously generates features for emotion-cause pairs and clauses,enhancing feature interactions between emotion clauses,cause clauses,and emotion-cause pairs.Second,the task alignment technique is applied to reduce the labeling distance between emotion-cause pair extraction and the two assistant tasks,capturing deep semantic information interactions among tasks.The proposed method is evaluated on a Chinese benchmark corpus using 10-fold cross-validation,assessing key performance metrics such as precision,recall,and F1 score.Experimental results demonstrate that the model achieves an F1 score of 76.05%,surpassing the state-of-the-art by 1.03%.The proposed model exhibits significant improvements in emotion-cause pair extraction(ECPE)and cause extraction(CE)compared to existing methods,validating its effectiveness.This research introduces a novel approach based on joint feature encoding and task alignment mechanisms,contributing to advancements in emotion-cause pair extraction.However,the study’s limitation lies in the data sources,potentially restricting the generalizability of the findings.展开更多
Feature fusion is an important technique in medical image classification that can improve diagnostic accuracy by integrating complementary information from multiple sources.Recently,Deep Learning(DL)has been widely us...Feature fusion is an important technique in medical image classification that can improve diagnostic accuracy by integrating complementary information from multiple sources.Recently,Deep Learning(DL)has been widely used in pulmonary disease diagnosis,such as pneumonia and tuberculosis.However,traditional feature fusion methods often suffer from feature disparity,information loss,redundancy,and increased complexity,hindering the further extension of DL algorithms.To solve this problem,we propose a Graph-Convolution Fusion Network with Self-Supervised Feature Alignment(Self-FAGCFN)to address the limitations of traditional feature fusion methods in deep learning-based medical image classification for respiratory diseases such as pneumonia and tuberculosis.The network integrates Convolutional Neural Networks(CNNs)for robust feature extraction from two-dimensional grid structures and Graph Convolutional Networks(GCNs)within a Graph Neural Network branch to capture features based on graph structure,focusing on significant node representations.Additionally,an Attention-Embedding Ensemble Block is included to capture critical features from GCN outputs.To ensure effective feature alignment between pre-and post-fusion stages,we introduce a feature alignment loss that minimizes disparities.Moreover,to address the limitations of proposed methods,such as inappropriate centroid discrepancies during feature alignment and class imbalance in the dataset,we develop a Feature-Centroid Fusion(FCF)strategy and a Multi-Level Feature-Centroid Update(MLFCU)algorithm,respectively.Extensive experiments on public datasets LungVision and Chest-Xray demonstrate that the Self-FAGCFN model significantly outperforms existing methods in diagnosing pneumonia and tuberculosis,highlighting its potential for practical medical applications.展开更多
In the realm of data privacy protection,federated learning aims to collaboratively train a global model.However,heterogeneous data between clients presents challenges,often resulting in slow convergence and inadequate...In the realm of data privacy protection,federated learning aims to collaboratively train a global model.However,heterogeneous data between clients presents challenges,often resulting in slow convergence and inadequate accuracy of the global model.Utilizing shared feature representations alongside customized classifiers for individual clients emerges as a promising personalized solution.Nonetheless,previous research has frequently neglected the integration of global knowledge into local representation learning and the synergy between global and local classifiers,thereby limiting model performance.To tackle these issues,this study proposes a hierarchical optimization method for federated learning with feature alignment and the fusion of classification decisions(FedFCD).FedFCD regularizes the relationship between global and local feature representations to achieve alignment and incorporates decision information from the global classifier,facilitating the late fusion of decision outputs from both global and local classifiers.Additionally,FedFCD employs a hierarchical optimization strategy to flexibly optimize model parameters.Through experiments on the Fashion-MNIST,CIFAR-10 and CIFAR-100 datasets,we demonstrate the effectiveness and superiority of FedFCD.For instance,on the CIFAR-100 dataset,FedFCD exhibited a significant improvement in average test accuracy by 6.83%compared to four outstanding personalized federated learning approaches.Furthermore,extended experiments confirm the robustness of FedFCD across various hyperparameter values.展开更多
Audio-visual scene classification(AVSC)poses a formidable challenge owing to the intricate spatial-temporal relationships exhibited by audio-visual signals,coupled with the complex spatial patterns of objects and text...Audio-visual scene classification(AVSC)poses a formidable challenge owing to the intricate spatial-temporal relationships exhibited by audio-visual signals,coupled with the complex spatial patterns of objects and textures found in visual images.The focus of recent studies has predominantly revolved around extracting features from diverse neural network structures,inadvertently neglecting the acquisition of semantically meaningful regions and crucial components within audio-visual data.The authors present a feature pyramid attention network(FPANet)for audio-visual scene understanding,which extracts semantically significant characteristics from audio-visual data.The authors’approach builds multi-scale hierarchical features of sound spectrograms and visual images using a feature pyramid representation and localises the semantically relevant regions with a feature pyramid attention module(FPAM).A dimension alignment(DA)strategy is employed to align feature maps from multiple layers,a pyramid spatial attention(PSA)to spatially locate essential regions,and a pyramid channel attention(PCA)to pinpoint significant temporal frames.Experiments on visual scene classification(VSC),audio scene classification(ASC),and AVSC tasks demonstrate that FPANet achieves performance on par with state-of-the-art(SOTA)approaches,with a 95.9 F1-score on the ADVANCE dataset and a relative improvement of 28.8%.Visualisation results show that FPANet can prioritise semantically meaningful areas in audio-visual signals.展开更多
Dynamic sign language recognition holds significant importance, particularly with the application of deep learning to address its complexity. However, existing methods face several challenges. Firstly, recognizing dyn...Dynamic sign language recognition holds significant importance, particularly with the application of deep learning to address its complexity. However, existing methods face several challenges. Firstly, recognizing dynamic sign language requires identifying keyframes that best represent the signs, and missing these keyframes reduces accuracy. Secondly, some methods do not focus enough on hand regions, which are small within the overall frame, leading to information loss. To address these challenges, we propose a novel Video Transformer Attention-based Network (VTAN) for dynamic sign language recognition. Our approach prioritizes informative frames and hand regions effectively. To tackle the first issue, we designed a keyframe extraction module enhanced by a convolutional autoencoder, which focuses on selecting information-rich frames and eliminating redundant ones from the video sequences. For the second issue, we developed a soft attention-based transformer module that emphasizes extracting features from hand regions, ensuring that the network pays more attention to hand information within sequences. This dual-focus approach improves effective dynamic sign language recognition by addressing the key challenges of identifying critical frames and emphasizing hand regions. Experimental results on two public benchmark datasets demonstrate the effectiveness of our network, outperforming most of the typical methods in sign language recognition tasks.展开更多
To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities...To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities,this paper proposes a Multi-modal Pre-synergistic Entity Alignmentmodel based on Cross-modalMutual Information Strategy Optimization(MPSEA).The model first employs independent encoders to process multi-modal features,including text,images,and numerical values.Next,a multi-modal pre-synergistic fusion mechanism integrates graph structural and visual modal features into the textual modality as preparatory information.This pre-fusion strategy enables unified perception of heterogeneous modalities at the model’s initial stage,reducing discrepancies during the fusion process.Finally,using cross-modal deep perception reinforcement learning,the model achieves adaptive multilevel feature fusion between modalities,supporting learningmore effective alignment strategies.Extensive experiments on multiple public datasets show that the MPSEA method achieves gains of up to 7% in Hits@1 and 8.2% in MRR on the FBDB15K dataset,and up to 9.1% in Hits@1 and 7.7% in MRR on the FBYG15K dataset,compared to existing state-of-the-art methods.These results confirm the effectiveness of the proposed model.展开更多
Video classification is an important task in video understanding and plays a pivotal role in intelligent monitoring of information content.Most existing methods do not consider the multimodal nature of the video,and t...Video classification is an important task in video understanding and plays a pivotal role in intelligent monitoring of information content.Most existing methods do not consider the multimodal nature of the video,and the modality fusion approach tends to be too simple,often neglecting modality alignment before fusion.This research introduces a novel dual stream multimodal alignment and fusion network named DMAFNet for classifying short videos.The network uses two unimodal encoder modules to extract features within modalities and exploits a multimodal encoder module to learn interaction between modalities.To solve the modality alignment problem,contrastive learning is introduced between two unimodal encoder modules.Additionally,masked language modeling(MLM)and video text matching(VTM)auxiliary tasks are introduced to improve the interaction between video frames and text modalities through backpropagation of loss functions.Diverse experiments prove the efficiency of DMAFNet in multimodal video classification tasks.Compared with other two mainstream baselines,DMAFNet achieves the best results on the 2022 WeChat Big Data Challenge dataset.展开更多
The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddi...The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.展开更多
In this paper,we study the problem of domain adaptation,which is a crucial ingredient in transfer learning with two domains,that is,the source domain with labeled data and the target domain with none or few labels.Dom...In this paper,we study the problem of domain adaptation,which is a crucial ingredient in transfer learning with two domains,that is,the source domain with labeled data and the target domain with none or few labels.Domain adaptation aims to extract knowledge from the source domain to improve the performance of the learning task in the target domain.A popular approach to handle this problem is via adversarial training,which is explained by the H△H-distance theory.However,traditional adversarial network architectures just align the marginal feature distribution in the feature space.The alignment of class condition distribution is not guaranteed.Therefore,we proposed a novel method based on pseudo labels and the cluster assumption to avoid the incorrect class alignment in the feature space.The experiments demonstrate that our framework improves the accuracy on typical transfer learning tasks.展开更多
With the popularisation of intelligent power,power devices have different shapes,numbers and specifications.This means that the power data has distributional variability,the model learning process cannot achieve suffi...With the popularisation of intelligent power,power devices have different shapes,numbers and specifications.This means that the power data has distributional variability,the model learning process cannot achieve sufficient extraction of data features,which seriously affects the accuracy and performance of anomaly detection.Therefore,this paper proposes a deep learning-based anomaly detection model for power data,which integrates a data alignment enhancement technique based on random sampling and an adaptive feature fusion method leveraging dimension reduction.Aiming at the distribution variability of power data,this paper developed a sliding window-based data adjustment method for this model,which solves the problem of high-dimensional feature noise and low-dimensional missing data.To address the problem of insufficient feature fusion,an adaptive feature fusion method based on feature dimension reduction and dictionary learning is proposed to improve the anomaly data detection accuracy of the model.In order to verify the effectiveness of the proposed method,we conducted effectiveness comparisons through elimination experiments.The experimental results show that compared with the traditional anomaly detection methods,the method proposed in this paper not only has an advantage in model accuracy,but also reduces the amount of parameter calculation of the model in the process of feature matching and improves the detection speed.展开更多
Sequence-based protein tertiary structure prediction is of fundamental importance because the function of a protein ultimately depends on its 3 D structure.An accurate residue-residue contact map is one of the essenti...Sequence-based protein tertiary structure prediction is of fundamental importance because the function of a protein ultimately depends on its 3 D structure.An accurate residue-residue contact map is one of the essential elements for current ab initio prediction protocols of 3 D structure prediction.Recently,with the combination of deep learning and direct coupling techniques,the performance of residue contact prediction has achieved significant progress.However,a considerable number of current Deep-Learning(DL)-based prediction methods are usually time-consuming,mainly because they rely on different categories of data types and third-party programs.In this research,we transformed the complex biological problem into a pure computational problem through statistics and artificial intelligence.We have accordingly proposed a feature extraction method to obtain various categories of statistical information from only the multi-sequence alignment,followed by training a DL model for residue-residue contact prediction based on the massive statistical information.The proposed method is robust in terms of different test sets,showed high reliability on model confidence score,could obtain high computational efficiency and achieve comparable prediction precisions with DL methods that relying on multi-source inputs.展开更多
从单张RGB图像中实现双手的3D交互式网格重建是一项极具挑战性的任务。由于双手之间的相互遮挡以及局部外观相似性较高,导致部分特征提取不够准确,从而丢失了双手之间的交互信息并使重建的手部网格与输入图像出现不对齐等问题。为了解...从单张RGB图像中实现双手的3D交互式网格重建是一项极具挑战性的任务。由于双手之间的相互遮挡以及局部外观相似性较高,导致部分特征提取不够准确,从而丢失了双手之间的交互信息并使重建的手部网格与输入图像出现不对齐等问题。为了解决上述问题,本文首先提出一种包含两个部分的特征交互适应模块,第一部分特征交互在保留左右手分离特征的同时生成两种新的特征表示,并通过交互注意力模块捕获双手的交互特征;第二部分特征适应则是将此交互特征利用交互注意力模块适应到每只手,为左右手特征注入全局上下文信息。其次,引入三层图卷积细化网络结构用于精确回归双手网格顶点,并通过基于注意力机制的特征对齐模块增强顶点特征和图像特征的对齐,从而增强重建的手部网格和输入图像的对齐。同时提出一种新的多层感知机结构,通过下采样和上采样操作学习多尺度特征信息。最后,设计相对偏移损失函数约束双手的空间关系。在InterHand2.6M数据集上的定量和定性实验表明,与现有的优秀方法相比,所提出的方法显著提升了模型性能,其中平均每关节位置误差(Mean Per Joint Position Error,MPJPE)和平均每顶点位置误差(Mean Per Vertex Position Error,MPVPE)分别降低至7.19 mm和7.33 mm。此外,在RGB2Hands和EgoHands数据集上进行泛化性实验,定性实验结果表明所提出的方法具有良好的泛化能力,能够适应不同环境背景下的手部网格重建。展开更多
文摘With the rapid expansion of social media,analyzing emotions and their causes in texts has gained significant importance.Emotion-cause pair extraction enables the identification of causal relationships between emotions and their triggers within a text,facilitating a deeper understanding of expressed sentiments and their underlying reasons.This comprehension is crucial for making informed strategic decisions in various business and societal contexts.However,recent research approaches employing multi-task learning frameworks for modeling often face challenges such as the inability to simultaneouslymodel extracted features and their interactions,or inconsistencies in label prediction between emotion-cause pair extraction and independent assistant tasks like emotion and cause extraction.To address these issues,this study proposes an emotion-cause pair extraction methodology that incorporates joint feature encoding and task alignment mechanisms.The model consists of two primary components:First,joint feature encoding simultaneously generates features for emotion-cause pairs and clauses,enhancing feature interactions between emotion clauses,cause clauses,and emotion-cause pairs.Second,the task alignment technique is applied to reduce the labeling distance between emotion-cause pair extraction and the two assistant tasks,capturing deep semantic information interactions among tasks.The proposed method is evaluated on a Chinese benchmark corpus using 10-fold cross-validation,assessing key performance metrics such as precision,recall,and F1 score.Experimental results demonstrate that the model achieves an F1 score of 76.05%,surpassing the state-of-the-art by 1.03%.The proposed model exhibits significant improvements in emotion-cause pair extraction(ECPE)and cause extraction(CE)compared to existing methods,validating its effectiveness.This research introduces a novel approach based on joint feature encoding and task alignment mechanisms,contributing to advancements in emotion-cause pair extraction.However,the study’s limitation lies in the data sources,potentially restricting the generalizability of the findings.
基金supported by the National Natural Science Foundation of China(62276092,62303167)the Postdoctoral Fellowship Program(Grade C)of China Postdoctoral Science Foundation(GZC20230707)+3 种基金the Key Science and Technology Program of Henan Province,China(242102211051,242102211042,212102310084)Key Scientiffc Research Projects of Colleges and Universities in Henan Province,China(25A520009)the China Postdoctoral Science Foundation(2024M760808)the Henan Province medical science and technology research plan joint construction project(LHGJ2024069).
文摘Feature fusion is an important technique in medical image classification that can improve diagnostic accuracy by integrating complementary information from multiple sources.Recently,Deep Learning(DL)has been widely used in pulmonary disease diagnosis,such as pneumonia and tuberculosis.However,traditional feature fusion methods often suffer from feature disparity,information loss,redundancy,and increased complexity,hindering the further extension of DL algorithms.To solve this problem,we propose a Graph-Convolution Fusion Network with Self-Supervised Feature Alignment(Self-FAGCFN)to address the limitations of traditional feature fusion methods in deep learning-based medical image classification for respiratory diseases such as pneumonia and tuberculosis.The network integrates Convolutional Neural Networks(CNNs)for robust feature extraction from two-dimensional grid structures and Graph Convolutional Networks(GCNs)within a Graph Neural Network branch to capture features based on graph structure,focusing on significant node representations.Additionally,an Attention-Embedding Ensemble Block is included to capture critical features from GCN outputs.To ensure effective feature alignment between pre-and post-fusion stages,we introduce a feature alignment loss that minimizes disparities.Moreover,to address the limitations of proposed methods,such as inappropriate centroid discrepancies during feature alignment and class imbalance in the dataset,we develop a Feature-Centroid Fusion(FCF)strategy and a Multi-Level Feature-Centroid Update(MLFCU)algorithm,respectively.Extensive experiments on public datasets LungVision and Chest-Xray demonstrate that the Self-FAGCFN model significantly outperforms existing methods in diagnosing pneumonia and tuberculosis,highlighting its potential for practical medical applications.
基金the National Natural Science Foundation of China(Grant No.62062001)Ningxia Youth Top Talent Project(2021).
文摘In the realm of data privacy protection,federated learning aims to collaboratively train a global model.However,heterogeneous data between clients presents challenges,often resulting in slow convergence and inadequate accuracy of the global model.Utilizing shared feature representations alongside customized classifiers for individual clients emerges as a promising personalized solution.Nonetheless,previous research has frequently neglected the integration of global knowledge into local representation learning and the synergy between global and local classifiers,thereby limiting model performance.To tackle these issues,this study proposes a hierarchical optimization method for federated learning with feature alignment and the fusion of classification decisions(FedFCD).FedFCD regularizes the relationship between global and local feature representations to achieve alignment and incorporates decision information from the global classifier,facilitating the late fusion of decision outputs from both global and local classifiers.Additionally,FedFCD employs a hierarchical optimization strategy to flexibly optimize model parameters.Through experiments on the Fashion-MNIST,CIFAR-10 and CIFAR-100 datasets,we demonstrate the effectiveness and superiority of FedFCD.For instance,on the CIFAR-100 dataset,FedFCD exhibited a significant improvement in average test accuracy by 6.83%compared to four outstanding personalized federated learning approaches.Furthermore,extended experiments confirm the robustness of FedFCD across various hyperparameter values.
基金Shenzhen Institute of Artificial Intelligence and Robotics for Society,Grant/Award Number:AC01202201003-02GuangDong Basic and Applied Basic Research Foundation,Grant/Award Number:2024A1515010252Longgang District Shenzhen's“Ten Action Plan”for Supporting Innovation Projects,Grant/Award Number:LGKCSDPT2024002。
文摘Audio-visual scene classification(AVSC)poses a formidable challenge owing to the intricate spatial-temporal relationships exhibited by audio-visual signals,coupled with the complex spatial patterns of objects and textures found in visual images.The focus of recent studies has predominantly revolved around extracting features from diverse neural network structures,inadvertently neglecting the acquisition of semantically meaningful regions and crucial components within audio-visual data.The authors present a feature pyramid attention network(FPANet)for audio-visual scene understanding,which extracts semantically significant characteristics from audio-visual data.The authors’approach builds multi-scale hierarchical features of sound spectrograms and visual images using a feature pyramid representation and localises the semantically relevant regions with a feature pyramid attention module(FPAM).A dimension alignment(DA)strategy is employed to align feature maps from multiple layers,a pyramid spatial attention(PSA)to spatially locate essential regions,and a pyramid channel attention(PCA)to pinpoint significant temporal frames.Experiments on visual scene classification(VSC),audio scene classification(ASC),and AVSC tasks demonstrate that FPANet achieves performance on par with state-of-the-art(SOTA)approaches,with a 95.9 F1-score on the ADVANCE dataset and a relative improvement of 28.8%.Visualisation results show that FPANet can prioritise semantically meaningful areas in audio-visual signals.
基金supported by the National Natural Science Foundation of China under Grant Nos.62076117 and 62166026the Jiangxi Provincial Key Laboratory of Virtual Reality under Grant No.2024SSY03151.
文摘Dynamic sign language recognition holds significant importance, particularly with the application of deep learning to address its complexity. However, existing methods face several challenges. Firstly, recognizing dynamic sign language requires identifying keyframes that best represent the signs, and missing these keyframes reduces accuracy. Secondly, some methods do not focus enough on hand regions, which are small within the overall frame, leading to information loss. To address these challenges, we propose a novel Video Transformer Attention-based Network (VTAN) for dynamic sign language recognition. Our approach prioritizes informative frames and hand regions effectively. To tackle the first issue, we designed a keyframe extraction module enhanced by a convolutional autoencoder, which focuses on selecting information-rich frames and eliminating redundant ones from the video sequences. For the second issue, we developed a soft attention-based transformer module that emphasizes extracting features from hand regions, ensuring that the network pays more attention to hand information within sequences. This dual-focus approach improves effective dynamic sign language recognition by addressing the key challenges of identifying critical frames and emphasizing hand regions. Experimental results on two public benchmark datasets demonstrate the effectiveness of our network, outperforming most of the typical methods in sign language recognition tasks.
基金partially supported by the National Natural Science Foundation of China under Grants 62471493 and 62402257(for conceptualization and investigation)partially supported by the Natural Science Foundation of Shandong Province,China under Grants ZR2023LZH017,ZR2024MF066,and 2023QF025(for formal analysis and validation)+1 种基金partially supported by the Open Foundation of Key Laboratory of Computing Power Network and Information Security,Ministry of Education,Qilu University of Technology(Shandong Academy of Sciences)under Grant 2023ZD010(for methodology and model design)partially supported by the Russian Science Foundation(RSF)Project under Grant 22-71-10095-P(for validation and results verification).
文摘To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities,this paper proposes a Multi-modal Pre-synergistic Entity Alignmentmodel based on Cross-modalMutual Information Strategy Optimization(MPSEA).The model first employs independent encoders to process multi-modal features,including text,images,and numerical values.Next,a multi-modal pre-synergistic fusion mechanism integrates graph structural and visual modal features into the textual modality as preparatory information.This pre-fusion strategy enables unified perception of heterogeneous modalities at the model’s initial stage,reducing discrepancies during the fusion process.Finally,using cross-modal deep perception reinforcement learning,the model achieves adaptive multilevel feature fusion between modalities,supporting learningmore effective alignment strategies.Extensive experiments on multiple public datasets show that the MPSEA method achieves gains of up to 7% in Hits@1 and 8.2% in MRR on the FBDB15K dataset,and up to 9.1% in Hits@1 and 7.7% in MRR on the FBYG15K dataset,compared to existing state-of-the-art methods.These results confirm the effectiveness of the proposed model.
基金Fundamental Research Funds for the Central Universities,China(No.2232021A-10)National Natural Science Foundation of China(No.61903078)+1 种基金Shanghai Sailing Program,China(No.22YF1401300)Natural Science Foundation of Shanghai,China(No.20ZR1400400)。
文摘Video classification is an important task in video understanding and plays a pivotal role in intelligent monitoring of information content.Most existing methods do not consider the multimodal nature of the video,and the modality fusion approach tends to be too simple,often neglecting modality alignment before fusion.This research introduces a novel dual stream multimodal alignment and fusion network named DMAFNet for classifying short videos.The network uses two unimodal encoder modules to extract features within modalities and exploits a multimodal encoder module to learn interaction between modalities.To solve the modality alignment problem,contrastive learning is introduced between two unimodal encoder modules.Additionally,masked language modeling(MLM)and video text matching(VTM)auxiliary tasks are introduced to improve the interaction between video frames and text modalities through backpropagation of loss functions.Diverse experiments prove the efficiency of DMAFNet in multimodal video classification tasks.Compared with other two mainstream baselines,DMAFNet achieves the best results on the 2022 WeChat Big Data Challenge dataset.
基金supported by National Natural Science Foundation of China(Grant No.51075323)
文摘The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.
基金supported by the National Key Research and Development Program of China(No.2016YFB0901902)the National Natural Science Foundation of China(No.61733018).
文摘In this paper,we study the problem of domain adaptation,which is a crucial ingredient in transfer learning with two domains,that is,the source domain with labeled data and the target domain with none or few labels.Domain adaptation aims to extract knowledge from the source domain to improve the performance of the learning task in the target domain.A popular approach to handle this problem is via adversarial training,which is explained by the H△H-distance theory.However,traditional adversarial network architectures just align the marginal feature distribution in the feature space.The alignment of class condition distribution is not guaranteed.Therefore,we proposed a novel method based on pseudo labels and the cluster assumption to avoid the incorrect class alignment in the feature space.The experiments demonstrate that our framework improves the accuracy on typical transfer learning tasks.
文摘With the popularisation of intelligent power,power devices have different shapes,numbers and specifications.This means that the power data has distributional variability,the model learning process cannot achieve sufficient extraction of data features,which seriously affects the accuracy and performance of anomaly detection.Therefore,this paper proposes a deep learning-based anomaly detection model for power data,which integrates a data alignment enhancement technique based on random sampling and an adaptive feature fusion method leveraging dimension reduction.Aiming at the distribution variability of power data,this paper developed a sliding window-based data adjustment method for this model,which solves the problem of high-dimensional feature noise and low-dimensional missing data.To address the problem of insufficient feature fusion,an adaptive feature fusion method based on feature dimension reduction and dictionary learning is proposed to improve the anomaly data detection accuracy of the model.In order to verify the effectiveness of the proposed method,we conducted effectiveness comparisons through elimination experiments.The experimental results show that compared with the traditional anomaly detection methods,the method proposed in this paper not only has an advantage in model accuracy,but also reduces the amount of parameter calculation of the model in the process of feature matching and improves the detection speed.
基金supported by the Strategic Priority CAS Project (No. XDB38050100)the National Key Research and Development Program of China (No. 2018YFB0204403)+4 种基金the National Natural Science Foundation of China (No. U1813203)the Shenzhen Basic Research Fund (Nos. RCYX2020071411473419,JCYJ20200109114818703,and JSGG20201102163800001)CAS Key Lab (No. 2011DP173015)Hong Kong Research Grant Council (No. GRF-17208019)the Outstanding Youth Innovation Fund (Doctoral Students) of CAS-SIAT (No. Y9G054)。
文摘Sequence-based protein tertiary structure prediction is of fundamental importance because the function of a protein ultimately depends on its 3 D structure.An accurate residue-residue contact map is one of the essential elements for current ab initio prediction protocols of 3 D structure prediction.Recently,with the combination of deep learning and direct coupling techniques,the performance of residue contact prediction has achieved significant progress.However,a considerable number of current Deep-Learning(DL)-based prediction methods are usually time-consuming,mainly because they rely on different categories of data types and third-party programs.In this research,we transformed the complex biological problem into a pure computational problem through statistics and artificial intelligence.We have accordingly proposed a feature extraction method to obtain various categories of statistical information from only the multi-sequence alignment,followed by training a DL model for residue-residue contact prediction based on the massive statistical information.The proposed method is robust in terms of different test sets,showed high reliability on model confidence score,could obtain high computational efficiency and achieve comparable prediction precisions with DL methods that relying on multi-source inputs.
文摘从单张RGB图像中实现双手的3D交互式网格重建是一项极具挑战性的任务。由于双手之间的相互遮挡以及局部外观相似性较高,导致部分特征提取不够准确,从而丢失了双手之间的交互信息并使重建的手部网格与输入图像出现不对齐等问题。为了解决上述问题,本文首先提出一种包含两个部分的特征交互适应模块,第一部分特征交互在保留左右手分离特征的同时生成两种新的特征表示,并通过交互注意力模块捕获双手的交互特征;第二部分特征适应则是将此交互特征利用交互注意力模块适应到每只手,为左右手特征注入全局上下文信息。其次,引入三层图卷积细化网络结构用于精确回归双手网格顶点,并通过基于注意力机制的特征对齐模块增强顶点特征和图像特征的对齐,从而增强重建的手部网格和输入图像的对齐。同时提出一种新的多层感知机结构,通过下采样和上采样操作学习多尺度特征信息。最后,设计相对偏移损失函数约束双手的空间关系。在InterHand2.6M数据集上的定量和定性实验表明,与现有的优秀方法相比,所提出的方法显著提升了模型性能,其中平均每关节位置误差(Mean Per Joint Position Error,MPJPE)和平均每顶点位置误差(Mean Per Vertex Position Error,MPVPE)分别降低至7.19 mm和7.33 mm。此外,在RGB2Hands和EgoHands数据集上进行泛化性实验,定性实验结果表明所提出的方法具有良好的泛化能力,能够适应不同环境背景下的手部网格重建。
