Abstract:Preface

Abstract:Recently, Graph Convolutional neural Networks (GCNs) have attracted much attention by generalizing convolutional neural networks to graph data, which includes redefining convolution and pooling operations on graphs. Due to the limitation that graph data can only focus on dyadic relations, it cannot perform well in real practice. In contrast, a hypergraph can capture high-order data interaction and is easy to deal with complex data representation using its flexible hyperedges. However, the existing methods for hypergraph convolutional networks are still not mature, and there is no effective operation for hypergraph pooling currently. Therefore, a hypergraph pooling network with a self-attention mechanism is proposed. Using a hypergraph structure for data modeling, this model can learn node hidden features with high-order data information through hypergraph convolution operation which introduces a self-attention mechanism, select important nodes both on structure and content through hypergraph pooling operation, and then obtain more accurate hypergraph representation. Experiments on text classification, dish classification, and protein classification tasks show that the proposed method outperforms recent state-of-the-art methods.

Abstract:Temporal Knowledge Graph (TKG) reasoning has attracted wide attention of researchers. Existing TKG reasoning methods have made great progress through modeling historical information. However, the time variability and unseen entities (relations) are still two major challenges that hinder the further improvement of this field. Moreover, since the structural information and temporal dependencies of the historical subgraph sequence have to be modeled, the traditional embedding-based methods often have high time consumption in the training and predicting processes, which greatly limits the application of the reasoning model in real-world scenarios. To address these issues, in this paper we propose a frequency statistical network for TKG reasoning, namely FS-Net. On the one hand, FS-Net continuously generates time-varying scores for the predictions at the changing timestamps based on the latest short-term historical fact frequency statistics; on the other hand, based on the fact frequency statistics at the current timestamp, FS-Net supplements the historical unseen entities (relations) for the predictions; in particular, FS-Net does not need training and has a very high time efficiency. Plenty of experiments on two TKG benchmark datasets demonstrate that FS-Net outperforms the baseline models.

Abstract:As a learning method of heterogeneous graph representation, heterogeneous graph neural networks can effectively extract complex structural and semantic information from heterogeneous graphs, and perform excellently in node classification and link prediction tasks to provide strong support for the representation and analysis of knowledge graphs. Due to the existence of some noisy interactions or missing interactions in the heterogeneous graphs, the heterogeneous graph neural network incorporates erroneous neighbor features, thus affecting the overall performance of the model. To solve the above problems, in this paper we proposes a heterogeneous graph structure learning model enhanced by multi-view contrast. Firstly, the semantic information in the heterogeneous graph is maintained by the meta-path, and the similarity graph is generated by calculating the feature similarity among the nodes under each meta-path, which is fused with the meta-path graph to optimize the graph structure. By contrasting the similarity graph and meta-path graph as multiple views, the graph structure is optimized without supervision information, and the dependence on supervision signals is eliminated. Finally, for addressing the problem that the learning ability of the neural network model is insufficient at the initial training stage and there are often erroneous interactions in the generated graph structure, we design a progressive graph structure fusion method. Through incremental weighted addition of meta-path graphs and similarity graphs, the weight of similarity graphs in the fusion is changed. This not only prevents erroneous interactions from being introduced in the initial training stage but also achieves the purpose of employing the interactions in similarity graphs to suppress interference interactions or complete missing interactions, which leads to the optimized heterogeneous structure. Meanwhile, node classification and node clustering are selected as the verification tasks of graph structure learning. The experimental results on four real heterogeneous graph datasets prove that the proposed learning method is feasible and effective. Compared with the optimal comparison model, the performance of this model has been significantly improved under both tasks.

Abstract:A knowledge hypergraph is a form of heterogeneous graph representing the real world through $n$-ary relations, but existing knowledge hypergraphs are usually incomplete in both general and vertical domains. Therefore, it is challenging to infer the missing links from the existing links in knowledge hypergraphs. Most of the current studies employ knowledge representation learning methods based on $n$-ary relations to accomplish link prediction in knowledge hypergraphs, but they only learn the embedding vectors of entities and relations from time-unknown hyperedges without considering the influence of temporal factors on the dynamic evolution of facts, which results in poor prediction performance in dynamic environments. Firstly, based on the definition of temporal knowledge hypergraphs proposed by this paper for the first time, this paper puts forward a link prediction model for temporal knowledge hypergraphs and learns static and dynamic representations of entities from their roles, positions, and timestamps of temporal hyperedges. Then these representations are merged in a certain proportion and utilized as final entity embedding vectors for link prediction tasks to realize the full exploitation of hyperedge temporal information. Meanwhile, it is theoretically proven that the proposed model is fully expressive with linear space complexity. Additionally, a temporal knowledge hypergraph dataset CB67 is constructed from the public business data of listed companies, and a large number of experimental evaluations are conducted on this dataset. The experimental results show that the proposed model can effectively perform link prediction tasks on the temporal knowledge hypergraph dataset.

Abstract:In recommender systems, the cold-start issue is challenging due to the lack of interactions between users or items. Such an issue can be alleviated via data-level and model-level strategies. Traditional data-level methods employ auxiliary information like feature information to enhance the learning of user and item embeddings. Recently, Heterogeneous Information Networks (HINs) have been incorporated into the recommender system as they provide more fruitful auxiliary information and meaningful semantics. However, these models are unable to capture the structural and semantic information comprehensively and neglect the unlabeled information of HINs during training. Model-level methods propose to apply the meta-learning framework which naturally fits into the cold-start issue, as it learns the prior knowledge from similar tasks and adapts to new tasks quickly with few labeled samples. Therefore, we propose a contrastive meta-learning framework on HINs named CM-HIN, which addresses the cold-start issue at both data level and model level. In particular, we explore meta-path and network schema views to describe the higher-order and local structural information of HINs. Within meta-path and network schema views, contrastive learning is adopted to mine the unlabeled information of HINs and incorporate these two views. Extensive experiments on three benchmark datasets demonstrate that CM-HIN outperforms all state-of-the-art baselines in three cold-start scenarios.