Embedding the Web: An Open Billion-Scale Repository for Knowledge Graph Embeddings

Entities

Knowledge Graphs can contain millions of entities. Entities (like persons, animals or films) have identifiers. Use the button below to get some examples out of available entities.

Embeddings available:

Embeddings

Embeddings can be computed through analysis of properties and relationships of entities. Embeddings are vectors with the desired property that close embeddings correspond to a semantic similarity of related entities. Use the form below to view an example.

Data

Embeddings computed on data from Web data commons with DeCal. You can access the underlying data dump here, which currently holds about 9 billion IRIs and is continuously updated.

Old data related to the article Universal Knowledge Graph Embeddings can be found on this page.

Collaboration

And now? We offer embeddings as a service with a convenient API.
Our experts are continuously working on updates, we try to hide them in the news.