Name
Abstract | ||
|---|---|---|
Representation learning of knowledge graph aims to transform both the entities and relations into continuous low-dimensional vector space. Though there have been a variety of models for knowledge graph embedding, most existing latent-based models merely explain triples via latent features, while supplementary rich inference patterns hidden in the observed graph features have not been fully employed. For this reason, in this paper we propose the discriminative path-based embedding model (DPTransE) which jointly learns from the latent features and graph features. Our model builds interactions between these two features, and uses the graph features as the crucial prior to offer precise and discriminative embedding. Experimental results demonstrate that our method outperforms other baselines on the task of link prediction and entity classification. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1007/978-3-319-76941-7_21 | ADVANCES IN INFORMATION RETRIEVAL (ECIR 2018) |
Keywords | Field | DocType |
Knowledge representation,Knowledge graph,Distributed representation | Graph,Data mining,Knowledge graph,Vector space,Knowledge representation and reasoning,Embedding,Computer science,Inference,Theoretical computer science,Discriminative model,Feature learning | Conference |
Volume | ISSN | Citations |
10772 | 0302-9743 | 2 |
PageRank | References | Authors |
0.37 | 18 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Maoyuan Zhang | 1 | 27 | 3.67 |
| Qi Wang | 2 | 2 | 0.37 |
| Wukui Xu | 3 | 2 | 0.37 |
| Wei Li | 4 | 4 | 1.42 |
| Shuyuan Sun | 5 | 2 | 0.37 |