Most existing network representation learning algorithms focus on network structures for learning.However,network structure is only one kind of view and feature for various networks,and it cannot fully reflect all cha...Most existing network representation learning algorithms focus on network structures for learning.However,network structure is only one kind of view and feature for various networks,and it cannot fully reflect all characteristics of networks.In fact,network vertices usually contain rich text information,which can be well utilized to learn text-enhanced network representations.Meanwhile,Matrix-Forest Index(MFI)has shown its high effectiveness and stability in link prediction tasks compared with other algorithms of link prediction.Both MFI and Inductive Matrix Completion(IMC)are not well applied with algorithmic frameworks of typical representation learning methods.Therefore,we proposed a novel semi-supervised algorithm,tri-party deep network representation learning using inductive matrix completion(TDNR).Based on inductive matrix completion algorithm,TDNR incorporates text features,the link certainty degrees of existing edges and the future link probabilities of non-existing edges into network representations.The experimental results demonstrated that TFNR outperforms other baselines on three real-world datasets.The visualizations of TDNR show that proposed algorithm is more discriminative than other unsupervised approaches.展开更多
Sparse representation has attracted extensive attention and performed well on image super-resolution（SR） in the last decade. However, many current image SR methods face the contradiction of detail recovery and artif...Sparse representation has attracted extensive attention and performed well on image super-resolution（SR） in the last decade. However, many current image SR methods face the contradiction of detail recovery and artifact suppression. We propose a multi-resolution dictionary learning（MRDL） model to solve this contradiction, and give a fast single image SR method based on the MRDL model. To obtain the MRDL model, we first extract multi-scale patches by using our proposed adaptive patch partition method（APPM）. The APPM divides images into patches of different sizes according to their detail richness. Then, the multiresolution dictionary pairs, which contain structural primitives of various resolutions, can be trained from these multi-scale patches.Owing to the MRDL strategy, our SR algorithm not only recovers details well, with less jag and noise, but also significantly improves the computational efficiency. Experimental results validate that our algorithm performs better than other SR methods in evaluation metrics and visual perception.展开更多
The encoding aperture snapshot spectral imaging system,based on the compressive sensing theory,can be regarded as an encoder,which can efficiently obtain compressed two-dimensional spectral data and then decode it int...The encoding aperture snapshot spectral imaging system,based on the compressive sensing theory,can be regarded as an encoder,which can efficiently obtain compressed two-dimensional spectral data and then decode it into three-dimensional spectral data through deep neural networks.However,training the deep neural net⁃works requires a large amount of clean data that is difficult to obtain.To address the problem of insufficient training data for deep neural networks,a self-supervised hyperspectral denoising neural network based on neighbor⁃hood sampling is proposed.This network is integrated into a deep plug-and-play framework to achieve self-supervised spectral reconstruction.The study also examines the impact of different noise degradation models on the fi⁃nal reconstruction quality.Experimental results demonstrate that the self-supervised learning method enhances the average peak signal-to-noise ratio by 1.18 dB and improves the structural similarity by 0.009 compared with the supervised learning method.Additionally,it achieves better visual reconstruction results.展开更多
Face recognition based on few training samples is a challenging task. In daily applications, sufficient training samples may not be obtained and most of the gained training samples are in various illuminations and pos...Face recognition based on few training samples is a challenging task. In daily applications, sufficient training samples may not be obtained and most of the gained training samples are in various illuminations and poses. Non-sufficient training samples could not effectively express various facial conditions, so the improvement of the face recognition rate under the non-sufficient training samples condition becomes a laborious mission. In our work, the facial pose pre-recognition(FPPR) model and the dualdictionary sparse representation classification(DD-SRC) are proposed for face recognition. The FPPR model is based on the facial geometric characteristic and machine learning, dividing a testing sample into full-face and profile. Different poses in a single dictionary are influenced by each other, which leads to a low face recognition rate. The DD-SRC contains two dictionaries, full-face dictionary and profile dictionary, and is able to reduce the interference. After FPPR, the sample is processed by the DD-SRC to find the most similar one in training samples. The experimental results show the performance of the proposed algorithm on olivetti research laboratory(ORL) and face recognition technology(FERET) databases, and also reflect comparisons with SRC, linear regression classification(LRC), and two-phase test sample sparse representation(TPTSSR).展开更多
Sparse-representation-based single-channel source separation,which aims to recover each source’s signal using its corresponding sub-dictionary,has attracted many scholars’attention.The basic premise of this model is...Sparse-representation-based single-channel source separation,which aims to recover each source’s signal using its corresponding sub-dictionary,has attracted many scholars’attention.The basic premise of this model is that each sub-dictionary possesses discriminative information about its corresponding source,and this information can be used to recover almost every sample from that source.However,in a more general sense,the samples from a source are composed not only of discriminative information but also common information shared with other sources.This paper proposes learning a discriminative high-fidelity dictionary to improve the separation performance.The innovations are threefold.Firstly,an extra sub-dictionary was combined into a conventional union dictionary to ensure that the source-specific sub-dictionaries can capture only the purely discriminative information for their corresponding sources because the common information is collected in the additional sub-dictionary.Secondly,a task-driven learning algorithm is designed to optimize the new union dictionary and a set of weights that indicate how much of the common information should be allocated to each source.Thirdly,a source separation scheme based on the learned dictionary is presented.Experimental results on a human speech dataset yield evidence that our algorithm can achieve better separation performance than either state-of-the-art or traditional algorithms.展开更多
Motivated by local coordinate coding(LCC) theory in nonlinear manifold learning, a new image representation model called local sparse representation(LSR) for astronomical image denoising was proposed. Borrowing ideas ...Motivated by local coordinate coding(LCC) theory in nonlinear manifold learning, a new image representation model called local sparse representation(LSR) for astronomical image denoising was proposed. Borrowing ideas from surrogate function and applying the iterative shrinkage-thresholding algorithm(ISTA), an iterative shrinkage operator for LSR was derived. Meanwhile, a fast approximated LSR method by first performing a K-nearest-neighbor search and then solving a l1optimization problem was presented under the guarantee of denoising performance. In addition, the LSR model and adaptive dictionary learning were incorporated into a unified optimization framework, which explicitly established the inner connection of them. Such processing allows us to simultaneously update sparse coding vectors and the dictionary by alternating optimization method. The experimental results show that the proposed method is superior to the traditional denoising method and reaches state-of-the-art performance on astronomical image.展开更多
针对基于图的无监督特征选择算法存在挖掘数据内在信息不充分,且易受噪声干扰难以获取更具有判别性特征的问题,提出一种基于广义不相关回归和潜在表示学习的无监督特征选择方法(uncorrelated regression and latent representation for ...针对基于图的无监督特征选择算法存在挖掘数据内在信息不充分,且易受噪声干扰难以获取更具有判别性特征的问题,提出一种基于广义不相关回归和潜在表示学习的无监督特征选择方法(uncorrelated regression and latent representation for unsupervised feature selection,URLUFS)。该方法将非负矩阵分解作用于广义不相关回归模型的投影矩阵,使投影矩阵实现非线性的维数约简并获得特征选择矩阵。在特征选择矩阵的基础上,引入自适应图学习来进一步挖掘数据的局部流形结构,并对特征选择矩阵施加范数约束以保持稀疏性。利用潜在表示对数据样本间的相互关系进行学习,引导回归模型中的伪标签矩阵,从而选择出更具有判别性的特征。在8个公开的数据集上进行了数值对比实验,实验结果表明:基于广义不相关回归和潜在表示学习的无监督特征选择算法明显优于其他8种无监督特征选择算法。展开更多
业务流程合规性检查可以帮助企业及早发现潜在问题,保证业务流程的正常运行和安全性。提出一种基于改进BERT(Bidirectional Encoder Representations from Transformers)和轻量化卷积神经网络(CNN)的业务流程合规性检查方法。首先,根据...业务流程合规性检查可以帮助企业及早发现潜在问题,保证业务流程的正常运行和安全性。提出一种基于改进BERT(Bidirectional Encoder Representations from Transformers)和轻量化卷积神经网络(CNN)的业务流程合规性检查方法。首先,根据历史事件日志中的轨迹提取轨迹前缀,构造带拟合情况标记的数据集;其次,使用融合相对上下文关系的BERT模型完成轨迹特征向量的表示;最后,使用轻量化CNN模型构建合规性检查分类器,完成在线业务流程合规性检查,有效提高合规性检查的准确率。在5个真实事件日志数据集上进行实验,结果表明,该方法相比Word2Vec+CNN模型、Transformer模型、BERT分类模型在准确率方面有较大提升,且与传统BERT+CNN相比,所提方法的准确率最高可提升2.61%。展开更多
基金Projects(11661069,61763041) supported by the National Natural Science Foundation of ChinaProject(IRT_15R40) supported by Changjiang Scholars and Innovative Research Team in University,ChinaProject(2017TS045) supported by the Fundamental Research Funds for the Central Universities,China
文摘Most existing network representation learning algorithms focus on network structures for learning.However,network structure is only one kind of view and feature for various networks,and it cannot fully reflect all characteristics of networks.In fact,network vertices usually contain rich text information,which can be well utilized to learn text-enhanced network representations.Meanwhile,Matrix-Forest Index(MFI)has shown its high effectiveness and stability in link prediction tasks compared with other algorithms of link prediction.Both MFI and Inductive Matrix Completion(IMC)are not well applied with algorithmic frameworks of typical representation learning methods.Therefore,we proposed a novel semi-supervised algorithm,tri-party deep network representation learning using inductive matrix completion(TDNR).Based on inductive matrix completion algorithm,TDNR incorporates text features,the link certainty degrees of existing edges and the future link probabilities of non-existing edges into network representations.The experimental results demonstrated that TFNR outperforms other baselines on three real-world datasets.The visualizations of TDNR show that proposed algorithm is more discriminative than other unsupervised approaches.
文摘Sparse representation has attracted extensive attention and performed well on image super-resolution（SR） in the last decade. However, many current image SR methods face the contradiction of detail recovery and artifact suppression. We propose a multi-resolution dictionary learning（MRDL） model to solve this contradiction, and give a fast single image SR method based on the MRDL model. To obtain the MRDL model, we first extract multi-scale patches by using our proposed adaptive patch partition method（APPM）. The APPM divides images into patches of different sizes according to their detail richness. Then, the multiresolution dictionary pairs, which contain structural primitives of various resolutions, can be trained from these multi-scale patches.Owing to the MRDL strategy, our SR algorithm not only recovers details well, with less jag and noise, but also significantly improves the computational efficiency. Experimental results validate that our algorithm performs better than other SR methods in evaluation metrics and visual perception.
基金Supported by the Zhejiang Provincial"Jianbing"and"Lingyan"R&D Programs(2023C03012,2024C01126)。
文摘The encoding aperture snapshot spectral imaging system,based on the compressive sensing theory,can be regarded as an encoder,which can efficiently obtain compressed two-dimensional spectral data and then decode it into three-dimensional spectral data through deep neural networks.However,training the deep neural net⁃works requires a large amount of clean data that is difficult to obtain.To address the problem of insufficient training data for deep neural networks,a self-supervised hyperspectral denoising neural network based on neighbor⁃hood sampling is proposed.This network is integrated into a deep plug-and-play framework to achieve self-supervised spectral reconstruction.The study also examines the impact of different noise degradation models on the fi⁃nal reconstruction quality.Experimental results demonstrate that the self-supervised learning method enhances the average peak signal-to-noise ratio by 1.18 dB and improves the structural similarity by 0.009 compared with the supervised learning method.Additionally,it achieves better visual reconstruction results.
基金supported by the National Natural Science Foundation of China(6137901061772421)
文摘Face recognition based on few training samples is a challenging task. In daily applications, sufficient training samples may not be obtained and most of the gained training samples are in various illuminations and poses. Non-sufficient training samples could not effectively express various facial conditions, so the improvement of the face recognition rate under the non-sufficient training samples condition becomes a laborious mission. In our work, the facial pose pre-recognition(FPPR) model and the dualdictionary sparse representation classification(DD-SRC) are proposed for face recognition. The FPPR model is based on the facial geometric characteristic and machine learning, dividing a testing sample into full-face and profile. Different poses in a single dictionary are influenced by each other, which leads to a low face recognition rate. The DD-SRC contains two dictionaries, full-face dictionary and profile dictionary, and is able to reduce the interference. After FPPR, the sample is processed by the DD-SRC to find the most similar one in training samples. The experimental results show the performance of the proposed algorithm on olivetti research laboratory(ORL) and face recognition technology(FERET) databases, and also reflect comparisons with SRC, linear regression classification(LRC), and two-phase test sample sparse representation(TPTSSR).
基金This work was supported by the National Natural Science Foundation of China(62001489)the scientific research planning project of National University of Defense Technology(JS19-04).
文摘Sparse-representation-based single-channel source separation,which aims to recover each source’s signal using its corresponding sub-dictionary,has attracted many scholars’attention.The basic premise of this model is that each sub-dictionary possesses discriminative information about its corresponding source,and this information can be used to recover almost every sample from that source.However,in a more general sense,the samples from a source are composed not only of discriminative information but also common information shared with other sources.This paper proposes learning a discriminative high-fidelity dictionary to improve the separation performance.The innovations are threefold.Firstly,an extra sub-dictionary was combined into a conventional union dictionary to ensure that the source-specific sub-dictionaries can capture only the purely discriminative information for their corresponding sources because the common information is collected in the additional sub-dictionary.Secondly,a task-driven learning algorithm is designed to optimize the new union dictionary and a set of weights that indicate how much of the common information should be allocated to each source.Thirdly,a source separation scheme based on the learned dictionary is presented.Experimental results on a human speech dataset yield evidence that our algorithm can achieve better separation performance than either state-of-the-art or traditional algorithms.
基金Project(60972114) supported by the National Natural Science Foundation of ChinaProject(2012M512168) supported by China Postdoctoral Science Foundation
文摘Motivated by local coordinate coding(LCC) theory in nonlinear manifold learning, a new image representation model called local sparse representation(LSR) for astronomical image denoising was proposed. Borrowing ideas from surrogate function and applying the iterative shrinkage-thresholding algorithm(ISTA), an iterative shrinkage operator for LSR was derived. Meanwhile, a fast approximated LSR method by first performing a K-nearest-neighbor search and then solving a l1optimization problem was presented under the guarantee of denoising performance. In addition, the LSR model and adaptive dictionary learning were incorporated into a unified optimization framework, which explicitly established the inner connection of them. Such processing allows us to simultaneously update sparse coding vectors and the dictionary by alternating optimization method. The experimental results show that the proposed method is superior to the traditional denoising method and reaches state-of-the-art performance on astronomical image.
文摘针对基于图的无监督特征选择算法存在挖掘数据内在信息不充分,且易受噪声干扰难以获取更具有判别性特征的问题,提出一种基于广义不相关回归和潜在表示学习的无监督特征选择方法(uncorrelated regression and latent representation for unsupervised feature selection,URLUFS)。该方法将非负矩阵分解作用于广义不相关回归模型的投影矩阵,使投影矩阵实现非线性的维数约简并获得特征选择矩阵。在特征选择矩阵的基础上,引入自适应图学习来进一步挖掘数据的局部流形结构,并对特征选择矩阵施加范数约束以保持稀疏性。利用潜在表示对数据样本间的相互关系进行学习,引导回归模型中的伪标签矩阵,从而选择出更具有判别性的特征。在8个公开的数据集上进行了数值对比实验,实验结果表明:基于广义不相关回归和潜在表示学习的无监督特征选择算法明显优于其他8种无监督特征选择算法。
文摘业务流程合规性检查可以帮助企业及早发现潜在问题,保证业务流程的正常运行和安全性。提出一种基于改进BERT(Bidirectional Encoder Representations from Transformers)和轻量化卷积神经网络(CNN)的业务流程合规性检查方法。首先,根据历史事件日志中的轨迹提取轨迹前缀,构造带拟合情况标记的数据集;其次,使用融合相对上下文关系的BERT模型完成轨迹特征向量的表示;最后,使用轻量化CNN模型构建合规性检查分类器,完成在线业务流程合规性检查,有效提高合规性检查的准确率。在5个真实事件日志数据集上进行实验,结果表明,该方法相比Word2Vec+CNN模型、Transformer模型、BERT分类模型在准确率方面有较大提升,且与传统BERT+CNN相比,所提方法的准确率最高可提升2.61%。
