Post on 21-Jan-2018
Joe PairmanLead Consultant, Mekon Ltd.
Andreas BlumauerCEO, Semantic Web Company
WEBINAR
Structured Content Meets Taxonomy
1
2AGENDA
Structured Content Meets Taxonomy
▸ Introduction▸ Why Embedded Semantics?▸ What is Structured Content?▸ Why does structured content need taxonomies &
semantic technologies?▸ Q&A Engines: Benefiting from Knowledge Graphs▸ Creating Knowledge Graphs with PoolParty 6.0
INTRODUCTIONMekon & Semantic Web Company:
When clients benefit from complementary partners
3
INTRODUCTION
4Semantic Web
Company
founder & CEO of
Andreas Blumauer
developer and vendor of
2004founded
6.0
current Version
active at
based on
Vienna
located
part ofTaxonomies Knowledge
Graphs
manages
standard for part of
is a
>200serves customers
Introduction: Joe Pairman
Was at HTC, doing:▸ Structured content for
user education:
▹ business case
▹ information model,
▹ CCMS
implementation
▹ Training
▸ Mobile content delivery
platform
▸ Localization / web
delivery
▸ Internal knowledge
management
Now at Mekon:Lead consultant:
▸ Structured content
consulting and coaching
▸ Requirements gathering
& solution design
▸ Pilots & proofs of
concept
▸ Applied taxonomy
5
EMBEDDED SEMANTICS
From Structured to Semantic Content Authoring
6
Fundamental assumption
People actually do not search for documents only,
they need facts and smaller chunks of information.
Machines should help to create links between those to give
answers to concrete questions.
7
8Benefiting from Semantics alongside the Content Life Cycle
Four-layered Content Architecture
9
And what’s it good for?
What is structured content?
Non-structured content• Whole pages, with a “body” blob
• Documents
• Publications
Structured content is made of sub-publication chunks• Sections with a heading
or sub-heading
• Paragraphs, tables,
figures
Structured content is made of semantic chunksSections, paragraphs, tables,
phrases, tagged according to the
type of information they contain
Structured content is made of semantic chunksMaybe
this:
Structured content is made of semantic chunksNot
this:
Consistent, predictable answer structure
Readers expect consistency
Structure helps authors
Structure helps authors
Other benefits: multichannel output
“Show Me” app on phone
Support content on desktop
PDF user guide
Other benefits: multichannel output
Other benefits: tailoring to context
?
Tailoring to context (needs more metadata)
http://mekon.poolparty.biz/Demoresetterms/ZoomBar
http://mekon.poolparty.biz/Demoresetterms/AutoFocusIndicator
http://mekon.poolparty.biz/Demoresetterms/ShakeDetectionIndicator
http://mekon.poolparty.biz/Demoresetterms/PhotoVideoModeSwitch
http://mekon.poolparty.biz/Demoresetterms/FlashButton
Tailoring to context (needs more metadata)
Tailoring to context (needs profiles or rules)
Not only filtering — you can also pull out relevant chunks of information independently
Shortdesc is a very important chunk of info~ SERP link previews tend to use it
~ It may improve ranking in the first place!
~ For site search, link previews are just as usefulApparently no shortdescs or equivalent
Shortdescs (as meta description & in page content)
Word choice is critical
Why does structured content need taxonomy & semantic technologies?
What are we still missing?Structured content helps us:
~ Formulate complete, well sequenced answers
~ Personalize the answers
~ Label individual snippets & facts within answers so they can be used independently
In itself, structured content doesn’t:
~ Provide a central location for all labels and names
~ Tell other machines what content is about in a way that they can use effectively
Taxonomies help us connect related labels (helping authors use relevant language)
Taxonomies relate concepts to each other
A concept is (minimum):
Concepts can tag whole topics/pages
https://mekon.poolparty.biz/mekonchef3/164
Each site/platform reads the taxonomy
https://mekon.poolparty.biz/mekonchef3/164
Filter results by:Preparation method Chop (23) Combine (2) Mince (3) Shaved ice (1) Shred (8)
Dietary suitability Gluten-free Halal▸ More…
Type of dish Main meal Side dish▸ More…
By using the URI, label changes are picked up
https://mekon.poolparty.biz/mekonchef3/164
Filter results by:Preparation method Chop (23) Combine (2) Mince (3) Shave (ice) (1) Shred (8)
Dietary suitability Gluten-free Halal▸ More…
Type of dish Main meal Side dish▸ More…
Hierarchy changes work too
https://mekon.poolparty.biz/mekonchef3/164
Filter results by:Preparation method Flavoring / tenderizing Marinate (5) Dry rub (3)
Food processing Chop (23) Combine (2) Mince (3) Shave (ice) (1) Shred (8)
Dietary suitability Gluten-free Halal▸ More…
Type of dish Main meal Side dish▸ More…
To get the most from structured content, tag chunks of content with concepts
Example from content modeling / taxonomy dev client
Taxonomy-sourced conditions: SKOS source
Taxonomy-sourced conditions: applied in editor
Taxonomy-sourced conditions: reader view
Taxonomy-sourced conditions: reader view
To get the most from structured content, tag inline entitiesContent marketing proof of concept, with input from:
Content Marketing proof of concept
Starting with a recipe taxonomy
Tech writers create tech docs
BENEFITING FROM KNOWLEDGE GRAPHS
Towards a Q&A engine based in knowledge graphs
49
50What’s the problem with Structured Content (like DITA), especially in large content repositories?
<party> <name>Social Democratic Party </name> <abstract> The Social Democratic Party of Germany is a social-democratic political party in Germany. The party, led by Chairman Martin Schulz since 2017, has become one of the two major contemporary political parties in Germany, along with the Christian Democratic Union (CDU). </abstract> <altname>SPD</altname> <founded>1863</founded> <country>Germany</country></party>
<mayor> <name>Michael Müller </name> <bio> Michael Müller (born 9 December 1964 in West Berlin) is a German politician of the SPD. He has been the Mayor of Berlin since 11 December 2014. </bio> <city>Berlin</city> <party>SPD</party> <year>2014</year> <profession>Printer</profession> <religion>Lutheran</religion></mayor>
▸ Use names for things consistently▸ Allow synonyms▸ Keep all names up to date▸ Include relations between things▸ Relate and link documents to
each other dynamically▸ Filter and aggregate documents,
also along hierarchies▸ Extract machine-processable
facts from document repositories▸ Do automatic content quality
checks▸ Provide answers to complex
questions
51Structured content can be the raw material for a knowledge base, help desk, or Q&A machines
Sadiq KhanSadiq Khan is the current Mayor of London since 2016. He is a lawyer and member of the Labour Party. Khan is a practicing Sunni Muslim.
Michael MüllerMichael Müller is a politician of the SPD. He has been the Mayor of Berlin since 2014. Müller is printer by profession and member of the Lutheran church.
Virginia RaggiVirginia Raggi is the current Mayor of Rome, who was elected in 2016. Raggi represents the Five Star Movement. She is Roman Catholic and studied law.
?Corine MauchCorine Mauch is a Swiss Social Democrat politician who serves as mayor of Zurich since 2009. Mauch studied agricultural economics.
Kadir TopbaşKadir Topbaş is a Turkish architect and the current mayor of Istanbul since 2004. He is member of AKP and is Sunni Muslim.
- Which mayors in capital cities of the EU belong to social democratic parties?
- Which mayors in European cities with more than 500,000 inhabitants are female?
- Which mayors have a background in social sciences?
- Which female mayors belong to a conservative party?
- Which mayors believe in Christianity, which believe in Islam?
- Which mayors belong to parties that have existed for more than 20 years?
52How to link (structured) content to background knowledge?
<mayor> <name>Sadiq Khan </name> <city>London</city> <party>Labour Party </party> <year>2016</year> <profession>Lawyer</profession> <religion>Sunni Islam </religion></mayor>
- Which mayors in capital cities of the EU belong to social democratic parties?
- Which city is a capital city?- Which city belongs to which country?- Which country belongs to the EU?- Which party is a social democratic party?
- Which mayors in European cities with more than 500,000 inhabitants are female?
- Which city is based in Europe?- How many inhabitants has a city?- Which mayor is male or female?
- Which mayors have a background in social sciences?- On which science is a profession based?
- Which female mayors belong to a conservative party?- Which mayor is male or female?- Which party is a conservative party?
- Which mayors believe in Christianity, which believe in Islam?
- Which religions are Christian?- Which religions are Muslim?
- Which mayors belong to parties that have existed for more than 20 years?
- When was a party founded?
<mayor> <name>Michael Müller </name> <city>Berlin</city> <party>SPD</party> <year>2014</year> <profession>Printer</profession> <religion>Lutheran</religion></mayor>
<mayor> <name>Virginia Raggi </name> <city>Rome</city> <party>Five Star Movement </party> <year>2016</year> <profession>Lawyer</profession> <religion>Catholic</religion></mayor>
<mayor> <name>Corine Mauch </name> <city>Zurich</city> <party>SP</party> <year>2009</year> <profession>Economist</profession> <religion>unknown</religion></mayor><mayor> <name>Kadir Topbaş </name> <city>Istanbul</city> <party>AKP</party> <year>2004</year> <profession>Architect</profession> <religion>Sunni Islam </religion></mayor>
53Semantic Web Approach
Knowledge Graphs based on standards and explicit semantics
Things and URIs instead of Strings and Literals
<mayor> <name>Michael Müller </name> <city>Berlin</city> <party>SPD</party> <year>2014</year> <profession>Printer</profession> <religion>Lutheran</religion></mayor>
@prefix taxo: <http://marklogic.poolparty.biz/politics/> .@prefix skos: <http://www.w3.org/2004/02/skos/core#> .@prefix onto: <http://marklogic.poolparty.biz/Politics-ontology/> .
taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 a onto:Politician .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 skos:prefLabel “Michael Müller” .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isLeaderOf taxo:0d1fe09c-6ee0-4af7-87f8-b62b2bfc7aa1 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isMemberOf taxo:7177cd40-a01c-426d-b622-7b687f98f413 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:hasProfession taxo:aacd-53a0-49f7-92b4-72b8ecb1139e .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:believesIn taxo:baeaa801-0829-478b-b969-d4350ec58c69 .
Michael Müller
SPD
Berlin Lutheran
Party
Politician
Place
Religion
54RDF as the base technology for Linked Data
<mayor> <name>Michael Müller </name> <city>Berlin</city> <party>SPD</party> <year>2014</year> <profession>Printer</profession> <religion>Lutheran</religion></mayor>
@prefix taxo: <http://marklogic.poolparty.biz/politics/> .@prefix skos: <http://www.w3.org/2004/02/skos/core#> .@prefix onto: <http://marklogic.poolparty.biz/Politics-ontology/> .
taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 a onto:Politician .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 skos:prefLabel “Michael Müller” .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isLeaderOf taxo:0d1fe09c-6ee0-4af7-87f8-b62b2bfc7aa1 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isMemberOf taxo:7177cd40-a01c-426d-b622-7b687f98f413 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:hasProfession taxo:aacd-53a0-49f7-92b4-72b8ecb1139e .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:believesIn taxo:baeaa801-0829-478b-b969-d4350ec58c69 .taxo:7177cd40-a01c-426d-b622-7b687f98f413 a onto:Party .taxo:7177cd40-a01c-426d-b622-7b687f98f413 skos:prefLabel “Social Democratic Party” .taxo:7177cd40-a01c-426d-b622-7b687f98f413 skos:altLabel “SPD” .taxo:7177cd40-a01c-426d-b622-7b687f98f413 onto:locatedIn taxo:fcc75dc0-eca4-4c04-a065-b6b2fb39329f .taxo:7177cd40-a01c-426d-b622-7b687f98f413 onto:foundedIn “1863” .
<party> <name>Social Democratic Party </name> <altname>SPD</altname> <founded>1863</founded> <country>Germany</country></party>
Michael Müller
SPD
Berlin Lutheran
Party
Politician
Place
Religion
Germany
1863
55Knowledge Graphs can be extended continuously and can be enriched by other Linked Data sources
<mayor> <name>Michael Müller </name> <city>Berlin</city> <party>SPD</party> <year>2014</year> <profession>Printer</profession> <religion>Lutheran</religion></mayor>
@prefix taxo: <http://marklogic.poolparty.biz/politics/> .@prefix skos: <http://www.w3.org/2004/02/skos/core#> .@prefix onto: <http://marklogic.poolparty.biz/Politics-ontology/> .
taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 a onto:Politician .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 skos:prefLabel “Michael Müller” .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isLeaderOf taxo:0d1fe09c-6ee0-4af7-87f8-b62b2bfc7aa1 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:isMemberOf taxo:7177cd40-a01c-426d-b622-7b687f98f413 .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:hasProfession taxo:aacd-53a0-49f7-92b4-72b8ecb1139e .taxo:2bde4d8c-e354-4b2b-a7d8-a488f6605b12 onto:believesIn taxo:baeaa801-0829-478b-b969-d4350ec58c69 .taxo:7177cd40-a01c-426d-b622-7b687f98f413 a onto:Party .taxo:7177cd40-a01c-426d-b622-7b687f98f413 skos:prefLabel “Social Democratic Party” .taxo:7177cd40-a01c-426d-b622-7b687f98f413 skos:altLabel “SPD” .taxo:7177cd40-a01c-426d-b622-7b687f98f413 onto:locatedIn taxo:fcc75dc0-eca4-4c04-a065-b6b2fb39329f .taxo:7177cd40-a01c-426d-b622-7b687f98f413 onto:foundedIn “1863” .
<party> <name>Social Democratic Party </name> <altname>SPD</altname> <founded>1863</founded> <country>Germany</country></party>
Michael Müller
SPD
Berlin Lutheran
Party
Politician
Place
Religion
Germany
Berlin3.6m
European Union
1863
56schema.org
57schema.org
CREATING KNOWLEDGE GRAPHS
Creation and Maintenance with PoolParty 6.0
58
59Create Semantics
PoolParty supports various methods to create consistent semantic metadata
PoolParty Thesaurus ServerTaxonomy, ontology, and
graph management; linked data harvesting, taxonomy linking
PoolParty PowerTaggingAutomatic content classification
and semantic enrichment
Corpus Analysis & Machine LearningFully automated generation of graph
providing language statistics and co-occurences
PoolParty Unified ViewsTriplification of structured and
semi-structured content; schema mapping, anything2RDF
Graphs + Machine Learning
PoolParty as a supervised learning system
60Content Manager
Integrator
Taxonomist/Ontologist
ThesaurusServer
Extractor
PowerTagging
uses API
is user of
is user of
is basis of
is basis of
Index
annotates
enriches
Corpus Learning/ Semantic Analysis
CMS
extends
is basis of
analyzesuses API
proposesextensions
Demo
Semi-automatic creation and maintenance of knowledge graphs with PoolParty
61
TWO INTEGRATIONSCENARIOS
62DAM/CMS
Option 1:Concepts are derived from taxonomy and tagging is stored together with the asset in the DAM/CMS
http://apple.com/macmini.jpg
http://apple.com/graph/1234PoolParty
API
Option 2:Concepts are derived from taxonomy, and tagging event is stored in a triple store by tying together assets with concepts from graph.
DAM/CMS
http://apple.com/macmini.jpg
http://apple.com/graph/1234PoolParty
API
http://apple.com/macmini.jpg
http://apple.com/macmini.jpg
http://apple.com/graph/1234
RDF Store
Wed 3 May, 2017User4711
DAM/CMS API
PoolParty
PoolParty
Beyond Semantic Search
Link unstructured information to semantic graphs with PoolParty
63
PoolParty GraphSearch
Linked Data Warehouse
Unstructured Information (e.g. SharePoint)
RDF TripleStore (Taxonomies,
Annotations and Graphs)
64Q&A and Graph-based Search with MarkLogic
Structured Content Audit & Semantic Web Starter Kit
65
GET STARTED
66
Get your test account atwww.poolparty.biz
PoolParty Academy
Get certified!
67
https://www.poolparty.biz/academy/
CONNECT
Joe PairmanLead Consultant, Mekon
▸ joe.pairman@mekon.com▸ http://www.mekon.com▸ https://www.linkedin.com/in/joepairman▸ https://twitter.com/joepairman
68
© Mekon - http://www.mekon.com
CONNECT
Andreas BlumauerCEO, Semantic Web Company
▸ andreas.blumauer@semantic-web.com▸ https://www.semantic-web.com ▸ https://www.linkedin.com/in/andreasblumauer▸ https://twitter.com/semwebcompany ▸ https://ablvienna.wordpress.com/
69
© Semantic Web Company - http://www.semantic-web.com and http://www.poolparty.biz/