单个较大非均匀超图聚类旨在将非均匀超图包含的节点划分为多个簇,使得同一簇内的节点更相似,而不同簇中的节点更不相似,具有广泛的应用场景。目前,最优的基于超图神经网络的非均匀超图聚类方法CIAH(co-cluster the interactions via at...单个较大非均匀超图聚类旨在将非均匀超图包含的节点划分为多个簇,使得同一簇内的节点更相似,而不同簇中的节点更不相似,具有广泛的应用场景。目前,最优的基于超图神经网络的非均匀超图聚类方法CIAH(co-cluster the interactions via attentive hypergraph neural network)虽然较好地学习了非均匀超图的关系信息,但仍存在两点不足:(1)对于局部关系信息的挖掘不足;(2)忽略了隐藏的高阶关系。因此,提出一种基于多尺度注意力和动态超图构建的非均匀超图聚类模型MADC(non-uniform hypergraph clustering combining multi-scale attention and dynamic construction)。一方面,使用多尺度注意力充分学习了超边中节点与节点之间的局部关系信息;另一方面,采用动态构建挖掘隐藏的高阶关系,进一步丰富了超图特征嵌入。真实数据集上的大量实验结果验证了MADC模型在非均匀超图聚类上的聚类准确率(accuracy,ACC)、标准互信息(normalized mutual information,NMI)和调整兰德指数(adjusted Rand index,ARI)均优于CIAH等所有Baseline方法。展开更多
知识超图(knowledge hypergraph,KHG)是一种超图结构的知识图谱。知识超图链接预测是基于已知的实体和关系来预测缺失的实体或关系,具有重要的意义和价值。然而,现有基于神经网络的知识超图链接预测方法,只关注关系事实局部的语义特征,...知识超图(knowledge hypergraph,KHG)是一种超图结构的知识图谱。知识超图链接预测是基于已知的实体和关系来预测缺失的实体或关系,具有重要的意义和价值。然而,现有基于神经网络的知识超图链接预测方法,只关注关系事实局部的语义特征,缺乏对关系事实之间关联特征的表示学习。针对以上问题,提出了一种基于图注意力网络与卷积神经网络的链接预测方法(knowledge prediction based on GAT and convolutional neural network,HPGC)。一方面,采用改进的卷积网络(convolutional neural network,CNN)提取知识超图中节点实体表示的局部特征;另一方面,使用改进的GAT对节点和关系进行注意力建模,捕获节点之间的全局特征关系,并将两者进行融合,从而获取关系事实更全面的邻域结构,丰富超图关系事实的语义表示。此外,针对HPGC的GAT层输出矢量问题,引入多层感知机(multilayer perceptron,MLP)和正则化技术,提高模型训练的泛化能力。真实数据集上的大量实验结果验证了所提出方法的预测性能均优于基线方法。展开更多
semantics information while maintaining spatial detail con-texts.Long-range context information plays a crucial role in this scenario.How-ever,the traditional convolution kernel only provides the local and small size ...semantics information while maintaining spatial detail con-texts.Long-range context information plays a crucial role in this scenario.How-ever,the traditional convolution kernel only provides the local and small size of the receptivefield.To address the problem,we propose a plug-and-play module aggregating both local and global information(aka LGIA module)to capture the high-order relationship between nodes that are far apart.We incorporate both local and global correlations into hypergraph which is able to capture high-order rela-tionships between nodes via the concept of a hyperedge connecting a subset of nodes.The local correlation considers neighborhood nodes that are spatially adja-cent and similar in the same CNN feature maps of magnetic resonance(MR)image;and the global correlation is searched from a batch of CNN feature maps of MR images in feature space.The influence of these two correlations on seman-tic segmentation is complementary.We validated our LGIA module on various CNN segmentation models with the cardiac MR images dataset.Experimental results demonstrate that our approach outperformed several baseline models.展开更多
文摘多行为推荐(multi-behavior recommendation,MBR)在互联网平台中愈发重要,但现有方法仍面临两大挑战:a)无法刻画用户不同行为下的复杂兴趣偏好;b)难以建模不同行为间的相互关系。基于此,提出一种对比学习增强的多行为超图神经网络模型(multi-behavior hypergraph neural network model enhanced with contrastive lear-ning,MBHCL),在建模用户复杂多类型交互的同时,结合对比学习捕获行为间共性与差异,以获取更优嵌入表示,缓解冷启动与数据稀疏问题。具体地,MBHCL首先构建用户-项目多行为交互超图,以刻画用户对项目不同维度的偏好;其次设计三个对比任务整合单行为表示,通过捕捉行为间的共性与差异获取全面用户兴趣偏好。最终,MBHCL在四个真实场景数据集上进行对比实验。结果表明,在Tmall和BeiBei数据集上,HIT和NDCG指标有至少4.8%的提升,在Kuairand和Yelp数据集上,HIT和NDCG指标至少提升3.6%,并通过消融实验验证了各模块的有效性,同时显著改善了冷启动用户推荐效果。
文摘知识超图(knowledge hypergraph,KHG)是一种超图结构的知识图谱。知识超图链接预测是基于已知的实体和关系来预测缺失的实体或关系,具有重要的意义和价值。然而,现有基于神经网络的知识超图链接预测方法,只关注关系事实局部的语义特征,缺乏对关系事实之间关联特征的表示学习。针对以上问题,提出了一种基于图注意力网络与卷积神经网络的链接预测方法(knowledge prediction based on GAT and convolutional neural network,HPGC)。一方面,采用改进的卷积网络(convolutional neural network,CNN)提取知识超图中节点实体表示的局部特征;另一方面,使用改进的GAT对节点和关系进行注意力建模,捕获节点之间的全局特征关系,并将两者进行融合,从而获取关系事实更全面的邻域结构,丰富超图关系事实的语义表示。此外,针对HPGC的GAT层输出矢量问题,引入多层感知机(multilayer perceptron,MLP)和正则化技术,提高模型训练的泛化能力。真实数据集上的大量实验结果验证了所提出方法的预测性能均优于基线方法。
基金supported by the Sichuan Science and Technology Program(Grant No.2019ZDZX0005,2019YFG0496,2020YFG0143,2019JDJQ0002 and 2020YFG0009).
文摘semantics information while maintaining spatial detail con-texts.Long-range context information plays a crucial role in this scenario.How-ever,the traditional convolution kernel only provides the local and small size of the receptivefield.To address the problem,we propose a plug-and-play module aggregating both local and global information(aka LGIA module)to capture the high-order relationship between nodes that are far apart.We incorporate both local and global correlations into hypergraph which is able to capture high-order rela-tionships between nodes via the concept of a hyperedge connecting a subset of nodes.The local correlation considers neighborhood nodes that are spatially adja-cent and similar in the same CNN feature maps of magnetic resonance(MR)image;and the global correlation is searched from a batch of CNN feature maps of MR images in feature space.The influence of these two correlations on seman-tic segmentation is complementary.We validated our LGIA module on various CNN segmentation models with the cardiac MR images dataset.Experimental results demonstrate that our approach outperformed several baseline models.
