๐ About the research: PyTorch-BigGraph is a project by Facebook Research and a "distributed system for learning graph embeddings for large graphs" which -in turn- is based on the PyTorch-BigGraph: A Large-scale Graph Embedding Framework paper.
๐พ About the dataset: The complete BigGraph Wikidata-dataset for Weaviate and how to import it can be found here on Github.
๐ About the vector search engine: Weaviate is an open source vector search engine used to render the results based on the dataset mentioned above.
๐ป About the demo: In this demo, you can search through the entire datasetโ (78M objects and vectors)โ by selecting a Wikidata Entity ID from the Wikidata set and searching for it on the right top of this page. The service does a similarity search in Weaviate and collects the images from the Wikidata API. It will also show the relevant GraphQL queries where possible. You can also run the setup yourself via this Github repo.
โญ If you like what you see, a +1 on the Weaviate Github repo is appreciated.