2020
DOI: 10.1609/aaai.v34i01.5330
|View full text |Cite
|
Sign up to set email alerts
|

Revisiting Graph Based Collaborative Filtering: A Linear Residual Graph Convolutional Network Approach

Abstract: Graph Convolutional Networks~(GCNs) are state-of-the-art graph based representation learning models by iteratively stacking multiple layers of convolution aggregation operations and non-linear activation operations. Recently, in Collaborative Filtering~(CF) based Recommender Systems~(RS), by treating the user-item interaction behavior as a bipartite graph, some researchers model higher-layer collaborative signals with GCNs. These GCN based recommender models show superior performance compared to traditional wo… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
248
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
3
2
2

Relationship

1
6

Authors

Journals

citations
Cited by 446 publications
(248 citation statements)
references
References 17 publications
0
248
0
Order By: Relevance
“…But different from LightGCN, we retain the inner product operation, whose effectiveness has been verified by experiments in NGCF. Moreover, we use the same layer fusion operation as that in NGCF, which, according to [4], is equivalent to using residual prediction. Specifically, for a connected user-item pair ( , ) in domain A, we define the propagation function of 's and 's features as follows,…”
Section: Feature Propagationmentioning
confidence: 99%
See 1 more Smart Citation
“…But different from LightGCN, we retain the inner product operation, whose effectiveness has been verified by experiments in NGCF. Moreover, we use the same layer fusion operation as that in NGCF, which, according to [4], is equivalent to using residual prediction. Specifically, for a connected user-item pair ( , ) in domain A, we define the propagation function of 's and 's features as follows,…”
Section: Feature Propagationmentioning
confidence: 99%
“…In recent years, inspired by the success of graph convolutional networks (GCN) in effectively extracting features in non-Euclidean spaces, some researchers try to exploit the user-item bipartite graph structure by propagating embeddings on it, aiming at achieving more effective embeddings [1,4,14,31]. For example, Wang et al [29] proposed NGCF, which follows the same propagation rules as in GCN (including feature transformation, neighborhood aggregation and nonlinear activation) to capture the high-order connectivity between users and items by stacking multiple feature propagation layers, and achieves promising results.…”
Section: Introductionmentioning
confidence: 99%
“…For plain text, we can segment words by spaces. As for the TeX formulas, we develop a TeX parsing tool 5 which treats the TeX commands as special words. Then we clean the word sequences, e.g.…”
Section: Exercise Preprocessingmentioning
confidence: 99%
“…where SP and DP are respectively the sets of similar pairs and dissimilar pairs, S • is the similarity score S v or S s , and δ (x) is an indicator function that returns 1 iff x is true. The ranking performance is evaluated with the widely used metric NDCG@K [5,35]:…”
Section: Evaluation Metricsmentioning
confidence: 99%
See 1 more Smart Citation