Content Understanding
Creating a machine understanding of content.
We've built domain specific ML models that extract meaning from documents (including text, images, and videos).
Using these models, documents are converted into neural embeddings, allowing machines to compare one item to another. We then tag content with entities from the knowledge graph, giving users an simple way of expressing their preferences.
These embeddings also allow us to compare documents to users, and perform search queries that go beyond naive string matching.