5. The Metamorphosis is one from
indexing a Web of URLs to indexing a
Web of Named Entities
6.
7. Why Change?
• The Inventor of the Web Updated it
• Microsoft was going there
• Google wanted to answer questions
• Yahoo introduced a Web of Concepts
• People search for Named Entities
15. Hello Direct Answers
“Now we are increasingly able to provide
direct answers -- even if you’re speaking
your question using Voice Search --
which makes it quicker, easier and more
natural to find what you’re looking for.”
Google FORM 10-K for 2014 (Fiscal Year December 31, 2014)
16. Direct Answers
“In each case we’re trying to get you
direct answers to your queries because
it’s quicker and less hassle than the
ten blue links Google used to show.”
We built Google for users, not websites
Eric Schmidt – Google Europe Blog 9/6/2014
18. Search engines like Google, Yahoo! and Bing have
already started displaying richer information for
some search queries, including maps and weather
(for location searches), reviews and prices (for
product search queries), and profiles (for people
searches).
A path already started?
19. What is a Named Entity?
Specific People, Places, and Things
including brands (Thanks, Rand!)
And Entities show up in Queries!
20. According to an internal study of Microsoft, at least
20-30% of queries submitted to Bing search are
simply name entities, and it is reported 71% of
queries contain name entities.
~ Building Taxonomy of Web Search Intents for
Name Entity Queries
21. Google Introduced a Knowledge
Graph filled with Named Entities
(May 16, 2012)
27. Best Answers to Fact Queries
Selecting the best answer to a
fact query from among a set
of potential answers
Invented by: Douglas L. T.
Rohde, Thomas W. Ritchford
Assignee: Google Inc.
US Patent 7,953,720
Filed: March 31, 2005
Granted: May 31, 2011
28. Fact Queries
• A Fact Repository is searched to find
an answer
• Potential Answers are scored based
upon confidence and importance.
• The Best scored answer is displayed.
31. Answer Box Snippets
Determination of a desired
repository
Invented by Michael Angelo,
David Braginsky, Jeremy
Ginsberg, and Simon Tong
US Patent Application
20070005568
Published January 4, 2007
32. Answer Box Snippets
• Google may display an answer box from Different
Repositories: Local, News, Books, Images, Shopping
• They fulfill Universal Search Results
• At 2006 SES, were referred to as “Vertical Creep
into Search Results” and “Answer Box” Results.
33.
34. Intent Queries
Natural Language Search Results for
Intent Queries
International Filing Date: 23.05.2014
Inventors: Tomer Shmiel, Dvir Keysar,
and Yonatan Erez
35. Natural Language Intent Queries
• Natural language Questions
• Answers from Authoritative Resources
• Answers from a Data Store of headers and answers (Like
in an FAQ)
• Answers follow an Intent Format; “what are the
symptoms of X”, “X Treatment”
• Answers fit into a results format like Search Result
Snippets
36.
37. Look at the sources…
http://www.webmd.com/a-to-
z-guides/dehydration-adults
38. Triggering Answer Boxes
Triggering Answer Boxes
Invented by Tal Cohen,
Ziv Bar-Yossef, Igor Tsvetkov,
Adi Mano, Oren Naim, Nitsan Oz,
Nir Andelman, Pravir K. Gupta
US Patent Application
20150169750
Published June 18, 2015
39. Triggering Answer Boxes
• Review Search Results for Any that might Trigger
Answer Box Results; Search Result for “Duke Blue
Devils” could be Football or Basketball, so answer
box would be, too.
• Review Topics Associated with Search Results (news,
sports, weather, television)
• These Answer box results might be personalized.
40.
41. Rich Content for Answers
Rich Content for Query Answers
Pub. No.: WO/2015/102869
Invented by: Gal Chechik, Eyal
Segalis, Yaniv Leviathan, and
Yoav Izur
International Filing Date: 15.12.2014
42. Rich Content for Answers
• Answers may go beyond Text to Images, Video,
Audio
• Rich Results may be based upon contextual queries
that combine the questions and answers
• An image result for “What is the capital of
California?” might use “capital of California,
Sacramento” as the query
43. Entity Triggering
Using an entity database to answer
entity-triggering questions
Inventors: Melissa K. Carroll, and
John J. Lee
US Patent 9,081,814 –
Granted July 14, 2015
44. Entity Triggering
• Similar to Fact Repository Approach
• Google looks for entities in queries.
• Google tries to including those entities in the
answers.
• “Identifying” the entities may help lead to answers
• Answers may be values of attributes of those
entities, such as, “What is the population of LA?”
45.
46. Combining Data & URLs
• Rich Snippets – Uses Data from Schema
Vocabulary
• Structured Snippets – Uses Data from
Tables
47. Data + URLs (Rich Snippets)
http://schema.org/
48. Rich Snippets
Generating specialized search results in response to patterned queries
Invented by Nicholas Brock Weininger, and Ramanathan V. Guha
US Patent 7,593,939 Granted September 22, 2009
53. Structured Snippets
Applying WebTables in Practice, by Sreeram
Balakrishnan, Alon Halevy, Boulos Harb, Hongrae
Lee, Jayant Madhavan, Afshin Rostamizadeh,
Warren Shen, Kenneth Wilder, Fei Wu, Cong Yu, 7th
Biennial Conference on Innovative Data Systems
Research (CIDR ’15) January 4-7, 2015,
Asilomar, California, USA
54. Structured Snippets
• Use Tables on your pages filled with data
• Use markup like Table Headings<th> for rows of the
tables
• Use Titles and captions descriptive of the Table’s
data
• Use Associated text on the Page the table is about
to help Google better understand it.
55. Further Reading on Google Answers
• Deep Learning - Teaching Machines to Read and
Comprehend (Deepmind)
• Information Extraction - Open Information
Extraction from the Web (Wavii)
• Tables - Applying WebTables in Practice (Alon
Halevy)
• Entities - Use These Tools To See What Entities Are
On A Web Page (Barbara Starr)
56. Questions
• Bill Slawski, Director of Search Marketing, Go Fish
Digital
• Editor of SEO by the Sea
• https://plus.google.com/u/0/+BillSlawski/posts
• https://twitter.com/bill_slawski
• https://www.facebook.com/bill.slawski