With the wide application of location-based social networks(LBSNs),personalized point of interest(POI)recommendation becomes popular,especially in the commercial field.Unfortunately,it is challenging to accurately rec...With the wide application of location-based social networks(LBSNs),personalized point of interest(POI)recommendation becomes popular,especially in the commercial field.Unfortunately,it is challenging to accurately recommend POIs to users because the user-POI matrix is extremely sparse.In addition,a user's check-in activities are affected by many influential factors.However,most of existing studies capture only few influential factors.It is hard for them to be extended to incorporate other heterogeneous information in a unified way.To address these problems,we propose a meta-path-based deep representation learning(MPDRL)model for personalized POI recommendation.In this model,we design eight types of meta-paths to fully utilize the rich heterogeneous information in LBSNs for the representations of users and POIs,and deeply mine the correlations between users and POIs.To further improve the recommendation performance,we design an attention-based long short-term memory(LSTM)network to learn the importance of different influential factors on a user's specific check-in activity.To verify the effectiveness of our proposed method,we conduct extensive experiments on a real-world dataset,Foursquare.Experimental results show that the MPDRL model improves at least 16.97%and 23.55%over all comparison methods in terms of the metric Precision@N(Pre@N)and Recall@N(Rec@N)respectively.展开更多
Information networks that can be extracted from many domains are widely studied recently. Different functions for mining these networks are proposed and developed, such as ranking, community detection, and link predic...Information networks that can be extracted from many domains are widely studied recently. Different functions for mining these networks are proposed and developed, such as ranking, community detection, and link prediction. Most existing network studies are on homogeneous networks, where nodes and links are assumed from one single type. In reality, however, heterogeneous information networks can better model the real-world systems, which are typically semi-structured and typed, following a network schema. In order to mine these heterogeneous information networks directly, we propose to explore the meta structure of the information network, i.e., the network schema. The concepts of meta-paths are proposed to systematically capture numerous semantic relationships across multiple types of objects, which are defined as a path over the graph of network schema. Meta-paths can provide guidance for search and mining of the network and help analyze and understand the semantic meaning of the objects and relations in the network. Under this framework, similarity search and other mining tasks such as relationship prediction and clustering can be addressed by systematic exploration of the network meta structure. Moreover, with user＇s guidance or feedback, we can select the best meta-path or their weighted combination for a specific mining task.展开更多
Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neu...Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neural networks to model inference.However,traditional knowledge graph are mostly concept-based,ignoring direct path evidence necessary for accurate reasoning.In this paper,we propose MRGNN(Meta-path Reasoning Graph Neural Network),a novel model that comprehensively captures sequential semantic information from concepts and paths.In MRGNN,meta-paths are introduced as direct inference evidence and an original graph neural network is adopted to aggregate features from both concepts and paths simultaneously.We conduct sufficient experiments on the CommonsenceQA and OpenBookQA datasets,showing the effectiveness of MRGNN.Also,we conduct further ablation experiments and explain the reasoning behavior through the case study.展开更多
异质图是由不同类型节点及边构成的图,可建模现实世界中各种类型对象及其关系。异质图嵌入旨在捕捉图中丰富的属性、结构和语义等信息,学习节点嵌入向量,用于节点分类、链接预测等任务,进而实现用户识别、商品推荐等应用。在异质图嵌入...异质图是由不同类型节点及边构成的图,可建模现实世界中各种类型对象及其关系。异质图嵌入旨在捕捉图中丰富的属性、结构和语义等信息,学习节点嵌入向量,用于节点分类、链接预测等任务,进而实现用户识别、商品推荐等应用。在异质图嵌入方法中,元路径通常被用来获取节点间的高阶结构和语义信息,然而现有方法忽略了元路径实例中不同类型节点或异质图中不同类型邻居节点的差异,导致信息丢失,进而影响节点嵌入质量。针对上述问题,提出基于数据增强的异质图注意力网络(Heterogeneous graph Attention Network based on Data Augmentation,HANDA),以更好地学习节点嵌入向量。首先,提出基于元路径邻居的边增强。该方法基于元路径获取节点的元路径邻居,用节点及其元路径邻居形成的语义边增强异质图。这些增强边不仅蕴含了节点间的高阶结构和语义,还缓解了异质图的稀疏性。其次,提出融入节点类型注意力的节点嵌入。该方法采用多头注意力从多个角度学习不同直接边邻居及增强边邻居的重要性并在注意力中融入节点的类型信息,进而通过消息传递、直接边邻居及增强边邻居同时获取节点的属性、高阶结构和语义信息,提升了节点嵌入质量。在真实数据集上的实验验证了HANDA模型在节点分类、链接预测任务上的效果优于基准模型。展开更多
异质图神经网络在挖掘复杂图数据任务中性能较优,但现有方法主要采用有监督学习范式,高度依赖节点标注信息,对原始图结构数据中的噪声链接较敏感,限制其在标注稀缺场景下的应用.针对上述问题,文中提出基于对比学习和结构更新机制的异质...异质图神经网络在挖掘复杂图数据任务中性能较优,但现有方法主要采用有监督学习范式,高度依赖节点标注信息,对原始图结构数据中的噪声链接较敏感,限制其在标注稀缺场景下的应用.针对上述问题,文中提出基于对比学习和结构更新机制的异质图结构学习方法(Heterogeneous Graph Structure Learning Based on Contrastive Learning and Structure Update Mechanism,HGSL-CL).首先,从原始数据中生成学习目标作为锚视图,结合类型感知特征映射与加权多视角相似度计算,生成学习者视图.然后,通过结构更新机制迭代优化锚视图,使用语义级注意力得到两个视角下的节点表示.最后,使用多层感知机将节点表示投影至同一维度空间,通过跨视角协同对比损失函数实现图结构优化,并引入融合节点拓扑相似度与属性相似度的正样本筛选策略,增强对比学习的判别能力.在3个数据集上的实验表明,HGSL-CL在节点分类、聚类等任务中性能较优,学习的图结构可泛化至半监督场景,取得比原始基线模型更优的性能,由此证实图结构学习的有效性.展开更多
跨域推荐技术通过深入挖掘及利用其他域的有用信息,有效提升目标域的推荐表现,为解决用户冷启动问题提供了一种有效途径。然而,当前跨域推荐方法存在局限,未能细粒度地扩展隐式关系,并且忽视了嵌入向量中可能包含的冗余信息,从而制约了...跨域推荐技术通过深入挖掘及利用其他域的有用信息,有效提升目标域的推荐表现,为解决用户冷启动问题提供了一种有效途径。然而,当前跨域推荐方法存在局限,未能细粒度地扩展隐式关系,并且忽视了嵌入向量中可能包含的冗余信息,从而制约了跨域推荐系统的性能。鉴于此,提出一种基于域内和域间元路径聚合的跨域推荐方法,IMCDR(intra-domain and inter-domain meta-paths aggregation based cross-domain recommendation)。IMCDR首先通过细粒度地计算实体多字段的语义嵌入,有效扩展用户-用户和物品-物品关系;然后,IMCDR基于域内元路径和域间元路径为每个节点分别生成私有特征和共享特征,并将它们有效融合,以获得更高质量的嵌入向量。在三个跨域推荐任务上的综合实验结果表明,IMCDR在有效性和性能上具有明显优势。展开更多
基金National Natural Science Foundation of China(No.61972080)Shanghai Rising-Star Program,China(No.19QA1400300)。
文摘With the wide application of location-based social networks(LBSNs),personalized point of interest(POI)recommendation becomes popular,especially in the commercial field.Unfortunately,it is challenging to accurately recommend POIs to users because the user-POI matrix is extremely sparse.In addition,a user's check-in activities are affected by many influential factors.However,most of existing studies capture only few influential factors.It is hard for them to be extended to incorporate other heterogeneous information in a unified way.To address these problems,we propose a meta-path-based deep representation learning(MPDRL)model for personalized POI recommendation.In this model,we design eight types of meta-paths to fully utilize the rich heterogeneous information in LBSNs for the representations of users and POIs,and deeply mine the correlations between users and POIs.To further improve the recommendation performance,we design an attention-based long short-term memory(LSTM)network to learn the importance of different influential factors on a user's specific check-in activity.To verify the effectiveness of our proposed method,we conduct extensive experiments on a real-world dataset,Foursquare.Experimental results show that the MPDRL model improves at least 16.97%and 23.55%over all comparison methods in terms of the metric Precision@N(Pre@N)and Recall@N(Rec@N)respectively.
基金supported in part by the U.S.Army Research Laboratory under Cooperative Agreement No.W911NF-09-2-0053(NS-CTA),NSF ⅡS-0905215,CNS-09-31975MIAS,a DHS-IDS Center for Multimodal Information Access and Synthesis at UIUC
文摘Information networks that can be extracted from many domains are widely studied recently. Different functions for mining these networks are proposed and developed, such as ranking, community detection, and link prediction. Most existing network studies are on homogeneous networks, where nodes and links are assumed from one single type. In reality, however, heterogeneous information networks can better model the real-world systems, which are typically semi-structured and typed, following a network schema. In order to mine these heterogeneous information networks directly, we propose to explore the meta structure of the information network, i.e., the network schema. The concepts of meta-paths are proposed to systematically capture numerous semantic relationships across multiple types of objects, which are defined as a path over the graph of network schema. Meta-paths can provide guidance for search and mining of the network and help analyze and understand the semantic meaning of the objects and relations in the network. Under this framework, similarity search and other mining tasks such as relationship prediction and clustering can be addressed by systematic exploration of the network meta structure. Moreover, with user＇s guidance or feedback, we can select the best meta-path or their weighted combination for a specific mining task.
基金supported by the Key Research and Development Program of Hubei Province(2020BAB017)the Scientific Research Center Program of National Language Commission(ZDI135-135)the Fundamental Research Funds for the Central Universities(KJ02502022-0155,CCNU22XJ037).
文摘Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neural networks to model inference.However,traditional knowledge graph are mostly concept-based,ignoring direct path evidence necessary for accurate reasoning.In this paper,we propose MRGNN(Meta-path Reasoning Graph Neural Network),a novel model that comprehensively captures sequential semantic information from concepts and paths.In MRGNN,meta-paths are introduced as direct inference evidence and an original graph neural network is adopted to aggregate features from both concepts and paths simultaneously.We conduct sufficient experiments on the CommonsenceQA and OpenBookQA datasets,showing the effectiveness of MRGNN.Also,we conduct further ablation experiments and explain the reasoning behavior through the case study.
文摘异质图是由不同类型节点及边构成的图,可建模现实世界中各种类型对象及其关系。异质图嵌入旨在捕捉图中丰富的属性、结构和语义等信息,学习节点嵌入向量,用于节点分类、链接预测等任务,进而实现用户识别、商品推荐等应用。在异质图嵌入方法中,元路径通常被用来获取节点间的高阶结构和语义信息,然而现有方法忽略了元路径实例中不同类型节点或异质图中不同类型邻居节点的差异,导致信息丢失,进而影响节点嵌入质量。针对上述问题,提出基于数据增强的异质图注意力网络(Heterogeneous graph Attention Network based on Data Augmentation,HANDA),以更好地学习节点嵌入向量。首先,提出基于元路径邻居的边增强。该方法基于元路径获取节点的元路径邻居,用节点及其元路径邻居形成的语义边增强异质图。这些增强边不仅蕴含了节点间的高阶结构和语义,还缓解了异质图的稀疏性。其次,提出融入节点类型注意力的节点嵌入。该方法采用多头注意力从多个角度学习不同直接边邻居及增强边邻居的重要性并在注意力中融入节点的类型信息,进而通过消息传递、直接边邻居及增强边邻居同时获取节点的属性、高阶结构和语义信息,提升了节点嵌入质量。在真实数据集上的实验验证了HANDA模型在节点分类、链接预测任务上的效果优于基准模型。
文摘异质图神经网络在挖掘复杂图数据任务中性能较优,但现有方法主要采用有监督学习范式,高度依赖节点标注信息,对原始图结构数据中的噪声链接较敏感,限制其在标注稀缺场景下的应用.针对上述问题,文中提出基于对比学习和结构更新机制的异质图结构学习方法(Heterogeneous Graph Structure Learning Based on Contrastive Learning and Structure Update Mechanism,HGSL-CL).首先,从原始数据中生成学习目标作为锚视图,结合类型感知特征映射与加权多视角相似度计算,生成学习者视图.然后,通过结构更新机制迭代优化锚视图,使用语义级注意力得到两个视角下的节点表示.最后,使用多层感知机将节点表示投影至同一维度空间,通过跨视角协同对比损失函数实现图结构优化,并引入融合节点拓扑相似度与属性相似度的正样本筛选策略,增强对比学习的判别能力.在3个数据集上的实验表明,HGSL-CL在节点分类、聚类等任务中性能较优,学习的图结构可泛化至半监督场景,取得比原始基线模型更优的性能,由此证实图结构学习的有效性.
文摘跨域推荐技术通过深入挖掘及利用其他域的有用信息,有效提升目标域的推荐表现,为解决用户冷启动问题提供了一种有效途径。然而,当前跨域推荐方法存在局限,未能细粒度地扩展隐式关系,并且忽视了嵌入向量中可能包含的冗余信息,从而制约了跨域推荐系统的性能。鉴于此,提出一种基于域内和域间元路径聚合的跨域推荐方法,IMCDR(intra-domain and inter-domain meta-paths aggregation based cross-domain recommendation)。IMCDR首先通过细粒度地计算实体多字段的语义嵌入,有效扩展用户-用户和物品-物品关系;然后,IMCDR基于域内元路径和域间元路径为每个节点分别生成私有特征和共享特征,并将它们有效融合,以获得更高质量的嵌入向量。在三个跨域推荐任务上的综合实验结果表明,IMCDR在有效性和性能上具有明显优势。
