Preview

This product is subject to the "Pre-GA Offerings Terms" in the General Service Terms section of theService Specific Terms. You can process personal data for this product as outlined in theCloud Data Processing Addendum, subject to the obligations and restrictions described in the agreement under which you access Google Cloud. Pre-GA products are available "as is" and might have limited support. For more information, see thelaunch stage descriptions.

Enterprise Knowledge Graph organizes siloedinformation into organizational knowledge, which involves consolidating, standardizing, and reconciling data in an efficient and useful way.

Entity Reconciliation API

Entity Reconciliation API is a lightweight, AI-powered, semantic clustering anddeduplication service for tabular data. It is a standalone API that wraps aroundthe Google core entity resolution engine. It helps customers to reconcile andjoin their data together, or to join their data with one or multiple third-party datasets.

Key features include:

• State-of-the-art quality pretrained model with Google data.
• Built-in entity enrichment, including geocoding.
• Google-scale clustering and reconciliation that handles a graph with a size up tobillions of nodes and trillions of edges.
• Stable Machine ID (MID) is a unique identifier for each of the entity clusters.
• Natively support up to 10 BigQuery tables.

How does it work

In relational databases, matching is primarily done through tightly managed keys, forexample, customer IDs. Some support fuzzy matching, for example, "RobertSmith" versus "Bob Smith". In Graphs, matching occurs through the use of any combination of fuzzytext, common relationships, entity types, and its attributes. This works wellwhen joining data across a heterogeneous mix of systems.

The Entity Reconciliation API reads data from source BigQuery tables given a set of entitytypes and files mapped to common schema. Then it performs knowledge extractionto turn the input relational data intoRDF triples like aknowledge graph representation. The Google Entity Reconciliation engine then builds a graph to cluster entities into different groups (entities in thesame group are considered as matched). The last step is to output thelinking result (matched and not-matched) to the customer-specified BigQuerytables in the form of a new unique identifier column (MID).

Google Knowledge Graph Search API

Search by keyword or lookup entities with an ID in the Google Knowledge Graph uses the Google Knowledge Graph Search API. The API uses standardschema.org types and is compliant with the JSON-LD specification.

Use the API to do the following:

• Perform entity resolution on an organization, a product, a location, a book, a movie, and many more entities.
• Link your private entity to a globally unique Cloud Knowledge Graph machine ID (MID).
• Get a ranked list of the most notable entities that match certain criteria.
• Annotate and organize content using the Google Knowledge Graph entities.
• Connect to the broader data ecosystem with additional IDs returned by the Google Knowledge Graph Search API, such as Google Place ID.
• Search for entities from MIDs returned by other Google Cloud APIs, such asCloud Vision API,Cloud Natural Language API andCloud Video Intelligence API

Google Knowledge Graph Search API editions

Google Knowledge Graph Search API offers two editions:

• Basic
• Advanced

Both editions support entity search and lookup. They also serve the same base knowledge graph entities built by Google Cloud. The Enterprise edition includes additional features such as linking to Google Place ID and additional data coverage.

If you're planning a new project, build your application with Cloud Knowledge GraphAdvanced edition to take advantage of new features and service improvements. TheBasic edition remains available, but doesn't support new features, high QPS, or additional security and compliance standards.

For a comparison between the two editions, seeCompare Basic and Advanced editions.