M2GRL: A Multi-task Multi-view Graph Representation Learning Framework for Web-scale Recommender Systems

Wang, Menghan; Lin, Yujie; Lin, Guli; Yang, Keping; Wu, Xiao-Ming

doi:10.1145/3394486.3403284

Cited by 48 publications

(17 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, Zheng et al [17] proposed the price-aware preference-modeling that employs the GCN to learn priceaware and category-dependent user representations. Wang et al [18] proposed a multi-task multi-view graph representation learning framework for web-scale recommender systems. Hu et al [19] proposed a graph neural news recommendation model based on unsupervised preference disentanglement.…”

Section: B the Gnn-based Item Recommendation Methodsmentioning

confidence: 99%

Neural Graph for Personalized Tag Recommendation

Chen

Zhang

et al. 2022

IEEE Intell. Syst.

View full text Add to dashboard Cite

Traditional personalized tag recommendation methods cannot guarantee that the collaborative signal hidden in the interactions among entities is effectively encoded in the process of learning the representations of entities, resulting in insufficient expressive capacity for characterizing the preferences or attributes of entities. In this paper, we firstly propose a graph neural networks boosted personalized tag recommendation model, namely NGTR, which integrates the graph neural networks into the pairwise interaction tensor factorization model. Specifically, we exploit the graph neural networks to capture the collaborative signal, and integrate the collaborative signal into the learning of representations of entities by transmitting and assembling the representations of neighbors along the interaction graphs. In addition, we also propose a light graph neural networks boosted personalized tag recommendation model, namely LNGTR. Different from NGTR, our proposed LNGTR model removes feature transformation and nonlinear activation components as well as adopts the weighted sum of the embeddings learned at all layers as the final embedding. Experimental results on real world datasets show that our proposed personalized tag recommendation models outperform the traditional tag recommendation methods.

show abstract

Section: B the Gnn-based Item Recommendation Methodsmentioning

confidence: 99%

Neural Graph for Personalized Tag Recommendation

Chen

Zhang

et al. 2022

IEEE Intell. Syst.

View full text Add to dashboard Cite

show abstract

“…We emphasize that UltraGCN is flexible to extend to model many different relation graphs, such as user-user graphs, itemitem graphs, and even knowlege graphs. In this work, we mainly demonstrate its use on the item-item co-occurrence graph, which has been shown to be useful for recommendation in [26]. We first build the item-item co-occurrence graph by linking items that have co-occurrences, which produces the following weighted adjacent matrix…”

Section: Learning On Item-item Graphmentioning

confidence: 99%

“…These GCN-based models are capable of exploiting higher-order connectivity between users and items, and therefore have achieved impressive performance gains for recommendation. PinSage [31] and M2GRL [26] are two successful use cases in industrial applications.…”

Section: Introductionmentioning

confidence: 99%

UltraGCN: Ultra Simplification of Graph Convolutional Networks for Recommendation

Mao¹,

Zhu²,

Xiao³

et al. 2021

Preprint

View full text Add to dashboard Cite

With the recent success of graph convolutional networks (GCNs), they have been widely applied for recommendation, and achieved impressive performance gains. The core of GCNs lies in its message passing mechanism to aggregate neighborhood information. However, we observed that message passing largely slows down the convergence of GCNs during training, especially for large-scale recommender systems, which hinders their wide adoption. LightGCN makes an early attempt to simplify GCNs for collaborative filtering by omitting feature transformations and nonlinear activations. In this paper, we take one step further to propose an ultra-simplified formulation of GCNs (dubbed UltraGCN), which skips infinite layers of message passing for efficient recommendation. Instead of explicit message passing, UltraGCN resorts to directly approximate the limit of infinite-layer graph convolutions via a constraint loss. Meanwhile, UltraGCN allows for more appropriate edge weight assignments and flexible adjustment of the relative importances among different types of relationships. This finally yields a simple yet effective UltraGCN model, which is easy to implement and efficient to train. Experimental results on four benchmark datasets show that UltraGCN not only outperforms the state-of-the-art GCN models but also achieves more than 10x speedup over LightGCN.

show abstract

“…Network embedding techniques (or network representation learning or graph embedding) utilize a dense low-dimensional vector to represent nodes [5][6][7]. This provides an efficient way to solve various graph analytic problems, including node classification [5][6][7], recommendation [8,9], link prediction [10,11]. Most existing network embedding techniques for node classification are designed for standard single-layer graph networks [1,2,5,[12][13][14], such as DeepWalk [13], node2vec [10], LINE [12], and classical graph neural networks (GNNs) such as graph convolutional networks (GCNs) [5], GAT Veličković and Cucurull [15], and GraphSAGE [14].…”

Section: Introductionmentioning

confidence: 99%

Node Classification in Attributed Multiplex Networks Using Random Walk and Graph Convolutional Networks

et al. 2022

View full text Add to dashboard Cite

Node classification, as a central task in the graph data analysis, has been studied extensively with network embedding technique for single-layer graph network. However, there are some obstacles when extending the single-layer network embedding technique to the attributed multiplex network. The classification of a given node in the attributed multiplex network must consider the network structure in different dimensions, as well as rich node attributes, and correlations among the different dimensions. Moreover, the distance node context information of a given node in each dimension will also affect the classification of the given node. In this study, a novel network embedding approach for the node classification of attributed multiplex networks using random walk and graph convolutional networks (AMRG) is proposed. A random walk network embedding technique was used to extract distant node information and the results are considered as pre-trained node features to be concatenated with the original node features inputted into the graph convolutional networks (GCNs) to learn node representations for each dimension. Besides, the consensus regularization is introduced to capture the similarities among different dimensions, and the learnable neural network parameters of GCNs for different dimensions are also constrained by the regularization mechanism to improve the correlations. As well as an attention mechanism is explored to infer the importance for a given node in different dimensions. Extensive experiments demonstrated that our proposed technique outperforms many competitive baselines on several real-world multiplex network datasets.

show abstract

M2GRL: A Multi-task Multi-view Graph Representation Learning Framework for Web-scale Recommender Systems

Cited by 48 publications

References 21 publications

Neural Graph for Personalized Tag Recommendation

Neural Graph for Personalized Tag Recommendation

UltraGCN: Ultra Simplification of Graph Convolutional Networks for Recommendation

Node Classification in Attributed Multiplex Networks Using Random Walk and Graph Convolutional Networks

Contact Info

Product

Resources

About