Matches in ScholarlyData for { <https://w3id.org/scholarlydata/inproceedings/eswc2016/paper/research/163> ?p ?o. }
Showing items 1 to 10 of
10
with 100 items per page.
- 163 creator cristobal-esteban.
- 163 creator volker-tresp.
- 163 creator yinchong-yang.
- 163 type InProceedings.
- 163 label "Embedding Mapping Approaches for Tensor Factorization and Knowledge Graph Modelling".
- 163 sameAs 163.
- 163 abstract "Latent embedding models, based for example on matrix and tensor factorization, are the basis of state-of-the art statistical solutions for modelling Knowledge Graphs and Recommender Systems to predict new links between known entities and/or relations. To be able to perform predictions for new entities and/or relation types, however, the model often has to be retrained completely to derive the new latent embeddings. This could be a potential limitation for data sets when fast predictions for new entities and relation types are required. In this paper we propose approaches that can map new entities into the existing latent embedding space learned from factorization models. Without retraining of any kind, our model is solely based on the observable ---even incomplete--- features of the new entities e.g. a subset of observed links to other known entities. We show that these mapping approaches are efficient and are applicable to a wide variety of existing factorization models, including nonlinear models. We perform experiments on multiple real-world datasets and evaluate the performances from different aspects.".
- 163 hasAuthorList authorList.
- 163 isPartOf proceedings.
- 163 title "Embedding Mapping Approaches for Tensor Factorization and Knowledge Graph Modelling".