Answering Queries with a Knowledge Graph

Answering Queries With A Knowledge Graph

by

Sharing is caring!

How Do Knowledge Graphs And Search Work Together?

There are at least a couple of processes that search and knowledge graphs can work together.

  1. Rewriting Queries – A search engine understands the data from apps on a mobile device and rewrites queries from that device using that data, which I explored in detail in the post User-Specific Knowledge Graphs to Support Queries and Predictions.
  2. Answering Queries With A Knowledge Graph – One of the best example of this, I wrote about in the post Ranked Entities in Search Results at Google

The Process Behind Answering Queries With A Knowledge Graph Has Been Granted by Google

The Ranked Entities post uses a patented process that describes how carousels of entities get created ranking those entities taken from SERPs that contain them that Google answered after turning those pages into a knowledge graph.

While there is a patent specifically about ranking entities in that manner, the part about turning SERPs onto knowledge graphs to answer queries is from a patent application that I wrote about in Answering Questions Using Knowledge Graphs. The patent is called “Natural Language Processing with an N-Gram Machine,” and this patent was granted Tuesday.

This patent is one of my favorite recent patent filings, and I know it is operating because it describes the process behind the Ranked Entities post that I wrote.

The granted version of the Answering Queries With A Knowledge Graph patent is at:

Natural language processing with an N-gram machine
Inventors: Ni Lao, Jiazhong Nie, and Fan Yang
Assignee: Google LLC
U.S. Patent: 11,256,866
Granted: February 22, 2022
Filed: October 25, 2017

popul;ar suspense films of 2021

Abstract

The present disclosure provides systems and methods that perform machine-learned natural language processing.

A computing system can include a machine-learned natural language processing model that consists of an encoder model trained to receive a natural language text body and output a knowledge graph and a programmer model trained to obtain a natural language question and output a program.

The computing system can include a computer-readable medium storing instructions that, when executed, cause the processor to perform operations.

The operations can include obtaining the natural language text body, inputting the natural language text body into the encoder model, receiving, as an output of the encoder model, the knowledge graph, obtaining the natural language question, inputting the natural language question into the programmer model, receiving the program as an output of the programmer model, and executing the program on the knowledge graph to produce an answer to the natural language question.

Sharing is caring!

5 thoughts on “Answering Queries With A Knowledge Graph”

  3. Thank you for the article. Bill, if the request does not contain a question, what principle is used to find the desired document? The knowledge graph is not used here? Thanks!

  4. Hi Viktor,

    A query in a search box at a search engine is usually implied to be a question by the search engine, even though it doesn’t use a question format. It is a query, or a keyword that a search engine will respond to. No need for question formatting. IN this case the knowledge graph can be used with these questions that are implied questions.

Comments are closed.