This article is about Google's specific implementation of knowledge graph technology. For knowledge engine technology in general, see
Knowledge engine.
Knowledge Graph data about Thomas Jefferson displayed on
Google Web Search, as of January 2015.
The
Knowledge Graph is a
knowledge base used by
Google to enhance its
search engine's search results with
semantic-search information gathered from a wide variety of sources. Knowledge Graph display was added to Google's search engine in 2012, starting in the United States, having been announced on May 16, 2012.
[1] It uses a
graph database to provide structured and detailed information about the topic in addition to a list of links to other sites. The goal is that users would be able to use this information to resolve their query without having to navigate to other sites and assemble the information themselves.
[2] The short summary provided in the knowledge graph is often used as a spoken answer in
Google Assistant searches.
[3]
According to some news websites, the implementation of Google's Knowledge Graph has played a role in the page view decline of various language
versions of Wikipedia.
[4][5][6][7] As of the end of 2016, knowledge graph holds over 70 billion facts.
[8]
According to Google, the information in the Knowledge Graph is derived from many sources, including the
CIA World Factbook,
Wikidata, and
Wikipedia.
[1] The feature is similar in intent to
answer engines such as
Wolfram Alpha and efforts such as
Linked Data and
DBpedia. As of 2012
[update], its
semantic network contained over 570 million objects and more than 18 billion facts about and relationships between different objects that are used to understand the meaning of the
keywords entered for the search.
[9][10]
On December 4, 2012, the Knowledge Graph was introduced in seven more languages: Spanish, French, German, Portuguese, Japanese, Russian, and Italian.
[11][12] During the Google I/O conference in May 2013, Google's
Amit Singhal presented on the future of search, explaining that a search engine's three primary functions will need to evolve and that search will need to: 1. Answer, 2. Converse, and 3. Anticipate.
[13] As part of his keynote talk Singhal asked: "A computer you can talk to? And it will answer everything you ask it?"
[14]
In August 2014, Google announced a new initiative, the
Knowledge Vault, which derives much of its data from the Knowledge Graph and the sources thereof, as well as harvesting its own data, ranking its reliability and compiling all results into a database of over 1.6 billion facts collected by
machine learning algorithms. On December 16, 2014, the
Freebase and Knowledge Graph team at Google announced that Freebase would shut-down in late 2015 and that they would help to transfer all of its data over to
Wikidata.
[15] In October 2016 Google announced that the Knowledge Graph now holds 70 billion facts.
[8]
Other companies' knowledge graphs:
- Microsoft Bing's Satori Knowledge Base, revealed to the public in mid-2013[16] (further details were not released)
- Yandex's Object Answer (ru), released in 2015[citation needed]
- Yahoo! and Baidu also have such technologies.[17][18]
- LinkedIn's Knowledge Graph, revealed to the public in Oct 2016.[19] It is a large knowledge base built upon LinkedIn "entities" such as 450M members, 190M historical jobs, 9M companies, 35K skills, 24K titles, 28K schools, 1.5K fields of study, 600+ degrees, 500+ certificates, 200+ countries, among other entities. It is a dynamic graph updated in real time upon member profile changes and when new entities emerge.
According to Google, information in the knowledge graph powers a "knowledge panel", which is a box containing information, and the box is presented at the top of search results.
[20] In May 2016,
The Washington Post reported that "knowledge panels and other sorts of 'rich answers' have mushroomed across Google, appearing atop the results on roughly one-third of its 100 billion monthly searches".
Dario Taraborelli of the
Wikimedia Foundation says that Google's omission of sources in its knowledge panels is designed so that the knowledge panel will seem more authoritative. The
Post reports that Google's knowledge panels are "frequently unattributed", such as a knowledge panel on the age of actress
Betty White which is "as unsourced and absolute as if handed down by God".
[21] Google will frequently scrape information from websites with varying degrees of success, but the increased use of
Schema markup means that information on websites can be more intelligently understood and utilized in the Knowledge Graph.
In
Google Assistant search results
[22] sources are included for search information in voice response and in the response card
Comments