In the Google patent “Providing Knowledge Panels With Search Results” is a reference to an earlier Google patent filing describing Knowledge Cards in depth. The patent provision is titled, “Apparatus and Method for Supplying Search Results with a knowledge Card”, and it is identified as being Patent Application No. 61/515,305, filed on Aug. 4, 2011.
This provisional patent is not linkable from the Web, otherwise, I would provide a link to it.
It is supposedly “incorporated fully” into that later patent filing, but a lot of details about what a knowledge card is have been left out of the later patent filing. I wrote about that later patent in a post titled, How Google Decides What to Know in Knowledge Graph Results, but the patent specifically about knowledge cards contains information not in the later patent.
Knowledge Panel results are part of Google’s Semantic Web search results which include a mix of result types such as Direct Answers, Structured Snippets, Rich Snippets and are part of an evolution of search results happening at Google and Bing and Microsoft that go much beyond yesterday’s 10-Blue links. I’ll be following this post with one about the rich search results that show up in response to queries at Bing.
The Knowledge Card patent tells us that, “This invention relates generally to user interfaces for presenting search results.”
Search results tend to require a searcher to review a number of snippets of information “for different links and/or clicking through to several websites.”
The purpose behind this patent involving knowledge cards is to provide a factual response to a query showing different aspects related to a “single conceptual entity.” It’s interesting that Google is using a Card type interface which seems to be more geared toward mobile search results than desktop results like Google Now Cards are.
Query Answers that provide knowledge cards are different than a Direct Answers to a natural language question, because They focus upon providing information about a specific entity. rather than an answer to a particular question.
What is contained in a knowledge card?
The provisional patent tells us that it is comprised of “condensed factual information that is
frequently sought by a user in association with a given query.” The factual information relates
to different aspects of that single conceptual entity associated with a query. It may contain the following components:
- facts and
- related searches
For example, a knowledge card for Abraham Lincoln contains his height, because he was the tallest US President, and a lot of people query his height.
The patent tells us that:
“The name is the most canonical descriptor of an entity. The name will usually be an alias for the entity that is mentioned most often as a title.” And it says that the name of an entity might be different than the name in a query The name will often be different than the name in a query, and provides and example of the query “Leonardo vinic” producing a knowledge card for “Leonardo da Vinci”. What is actually shown in response to that query is the choice of two different entities: one for “Leonardo da Vinci” and one for “Leonardo vinic”.
The patent tells us that the Knowledge card shouldn’t be distracting to searchers:
“The description should provide an adequate explanation of what the entity is without going into so much detail as to distract from the search results page.” Sources of Candidate descriptions might be chosen from “a variety of places, such as prefixes of text from trusted encyclopedia articles or top ranking web pages.”
Images shown for a knowledge panel might be taken from a “top ranking image for that entity from the search engine.”
Facts about the entity may also be displayed in the knowledge panel. For an individual, those facts may be specified as the date of birth, place of birth, date of death, place of death and nationality. Related searches may be shown looking a query log information.
How a knowledge Card Improves a search engine experience?
There are a number of ways a knowledge card is intended to help searchers.
- A knowledge card helps searchers with queries directed toward learning, browsing or discovery,
- It supplies users with basic factual information about a query
- It can help a user navigate to related content in a seamless and natural way
- A knowledge card supplies new content that may not otherwise be encountered
- A knowledge card helps users find information faster than they would by clicking through search results
- A knowledge card provides social proof that a site has a brand around it
- A Knowledge card can provide sentiment-based snippets of reviews
Knowledge Cards Used to Identify Entities
There are a couple of different ways that a knowledge card might be used. One of those is to identify a single conceptual entity, such as a “person, place, country, landmark, animal, historical event, organization, business or sports team.” It may also be used to “distinguish between distinct meanings associated with a query term.” For example, a query of “Phoenix” may produce a knowledge card with disambiguation information about the mythical bird and other disambiguation information about the capital city of Arizona. It may contain content that enables a searcher to choose between 2 different meanings.
Sources of Content for Cards
The kind of content shown may depend upon the type of entity. For example, a query involving a person could include a first set of standard information such as:
- birth place
- birth date
- career highlights
A query related to a place may include a second set of standard information, such as:
In both instances, we are often told that “Preferably, the knowledge card includes information from multiple sources. In this way, the knowledge card summarizes information from disparate sources.”
Factual Entities in Knowledge Cards
When a knowledge card might contain information about a factual entity, such as a person, place, country, landmark, animal, historical event, organization, business or sport team. a User interface processor might show off the content from that knowledge Card in a standard manner, using pre-existing content in a templated form.This is true for different types of entities.The patent provides examples of templates for:
Company Information Queries
Software Application Queries (games)
Highly Disambiguated Queries
For that last one, the example of “California universities” is provided, where there might be a number of possible results. When the patent was published, its authors mention listing a number of choices of universities, which Google was doing a few weeks ago, but is no longer doing now.