gentle introduction to semantic enrichment
DESCRIPTION
This talk is gentle introduction to the concept of semantic enrichment that demonstrates how publishers are using semantic technology such as Ontotext's GraphDB and publishing platform to make the most of their content.TRANSCRIPT
BSI Presentation, June 2014
A (gentle) Introduction to Semantic Enrichment
Jarred McGinnis, PhD
Some of the Problems in Publishing
#2June 2014
This is NOT an atomic
unit of information.
June 2014
The same word for different things.
June 2014
The Jordan problem
The different aspects of the same thing.
June 2014
The Schwarzenegger problem
The world is not static. Why is your content?
June 2014
The Prince problemThe RiM/Blackberry problemThe Yugoslavia problem
Semantics?
June 2014
Source: http://www.radarnetworks.com
Another development stage of the Internet.
June 2014
Computers are Stupid.
#9
“Nigel Farage”
June 2014
It’s all zeroes and ones.
#10
“Chris Huhne”
June 2014
June 2014
June 2014
June 2014
June 2014
Semantic Enrichment
June 2014
• Multi-staged text-analysis pipeline to enrich existing content.
• More than tags. Grounded in Semantics.
• Disambiguation (e.g. different people with the same name.)
Semantic Annotation
June 2014
This is creating context and from context...
#17June 2014
New information can be inferred.
#18June 2014
#19
New information can be inferred.
• A person is the president of a country.• Mr. Assad is the president of Syria.
• Syria is a country.• Assad is a person.
June 2014
Which leads to new information.
#20
• Would this chapter in this American textbook
satisfy the expected learning objectives for the UK?
Australia?
• What supplemental materials can we offer this
teacher given that a majority of her students are
struggling with Algebraic problems in her Level 1
physics?
• Which one of my authors have written the most
about ‘The Arab Spring’? What topics are my
authors writing about? Is there a gap in coverage?
What are my editors commissioning?
June 2014
And new products.
#21
• Celebrity X is in the news, trending on
twitter. What content do we have that
mentions X or is written by X? Automatically
generate a microsite on the imprint’s landing
page.
• Centennial of WW1. Automatically generated
map of Europe hyperlinked to chapters in a
book (or all the books in our catalogue).
Filter by campaign, general, belligerents or a
combination.
June 2014
Examples of Semantics in Use
June 2014
June 2014
E-commerce
June 2014
June 2014
SEO
Oct. 2014
Knowing What You Know
• Better Search
June 2014
June 2014
This slide should give a concrete example of one of our clients doing this.
June 2014
June 2014
June 2014
June 2014
June 2014
June 2014
June 2014
perform
comments
votes
posts
preview
read
contains leads to
readleads to
preview
Article
Search Action
Result
Date
FTS Q. Tag
Cat
Tag set
results
cattaxonomy
Search Log
-------------
-------------
-------------
-------------
-------------
June 2014
Making Use of What Others Know
• Making use of information and data from the ‘wilds’ of the internet safely.
June 2014
2011
Linked Data
June 2014
Automated Product Enrichment
June 2014
June 2014
Automated Product Enrichment
June 2014
June 2014
June 2014
Dynamic, finer grained management of content
June 2014
Athlete (10000+),Team (200+),Discipline (400+)and venue pages.
June 2014
June 2014
June 2014
Curation Becomes Smarter and Simpler
June 2014
An Example (Deckard Tool)
June 2014
June 2014
Thank [email protected]
June 2014