Making Sense of the Semantic Web
Nova Spivack CEO & Founder Radar Networks
Radar Networks 1
About This Talk
• Making sense of the semantic sector • Making the Semantic Web more useable • Future outlook • Twine.com
•Q & A
Radar Networks 2
The Big Opportunity…
The social graph just connects people The semantic graph connects everything…
People Companies Places Emails Products
Better search More targeted ads Smarter collaboration
Interests
Services
Deeper integration
Activities Projects Events Groups
Radar Networks 3
Web Pages
Richer content
Documents Multimedia
Better personalization
The third decade of the Web
• A period in time, not a technology… • Enrich the structure of the Web
Improve the quality of search, collaboration, publishing, advertising • Enables applications to become more integrated and intelligent
•
• Transform Web from fileserver to database
•
Semantic technologies will play a key role
Radar Networks 4
The Intelligence is in the Connections
Intelligent Web
Connections between Information
Web OS Semantic Web
Web 4.0
2020 - 2030 Intelligent personal agents
The Internet
FTP PC’s USENET
Distributed Search SWRL OWL 2010 - 2020 SPARQL Semantic Databases OpenID AJAX Semantic Search ATOM Widgets Social Web RSS Mashups P2P RDF Office 2.0 Javascript Flash SOAP XML 2000 - 2010 Weblogs Social Media Sharing Java The Web HTML SaaS Social Networking HTTP Directory Portals Wikis VR Keyword Search Lightweight Collaboration The PC BBS Gopher Websites 1990 - 2000 SQL MMO’s MacOS Groupware SGML Databases Windows File Servers
Web 3.0
Web 2.0
Web 1.0
IRC Email
PC Era
1980 - 1990
File Systems
Connections between people
Radar Networks 5
Beyond the Limits of Keyword Search
Productivity of Search
The Intelligent Web
Web 4.0
2020 - 2030 Reasoning
The Semantic Web
Web 3.0
2010 - 2020
Semantic Search
The Social Web The World Wide Web
Web2010 2.0 2000 Keyword search
Natural language search Tagging
Web2000 1.0 1990 The Desktop
PC Era
1980 - 1990
Databases
Directories
Files & Folders
Amount of data
Radar Networks 6
A Higher Resolution Web
Radar Networks 7
Five Approaches to Semantics
• Tagging
• Statistics • Linguistics • Semantic Web • Artificial Intelligence
Radar Networks 8
The Tagging Approach
• Pros
Easy for users to add and read tags • Tags are just strings • No algorithms or ontologies to deal with • No technology to learn
•
• Technorati • Del.icio.us
• Flickr
• Wikipedia
• Cons
Easy for users to add and read tags • Tags are just strings • No algorithms or ontologies to deal with • No technology to learn
•
Radar Networks 9
The Statistical Approach
• Pros:
Pure mathematical algorithms • Massively scaleable • Language independent
•
• Google • Lucene
• Cons:
No understanding of the content • Hard to craft good queries • Best for finding really popular things – not good at finding needles in haystacks • Not good for structured data
•
• Autonomy
Radar Networks 10
The Linguistic Approach
• Pros:
True language understanding • Extract knowledge from text • Best for search for particular facts or relationships • More precise queries
•
• Powerset • Hakia
• Inxight, Attensity, and others…
• Cons:
Computationally intensive • Difficult to scale • Lots of errors • Language-dependent
•
Radar Networks 11
The Semantic Web Approach
• Pros:
More precise queries • Smarter apps with less work • Not as computationally intensive • Share & link data between apps • Works for both unstructured and structured data
•
• Radar Networks • DBpedia Project
• Metaweb
• Cons:
Lack of tools • Difficult to scale • Who makes all the metadata?
•
Radar Networks 12
The Artificial Intelligence Approach
• Pros:
Smart in narrow domains • Answer questions intelligently • Reasoning and learning
•
• Cycorp
• Cons:
Computationally intensive • Difficult to scale • Extremely hard to program • Does not work well outside of narrow domains • Training takes a lot of work
•
Radar Networks 13
The Approaches Compared
Make the Data Smarter
A.I.
Semantic Web Linguistics
Tagging
Statistics
Make the software smarter
Radar Networks 14
Two Paths to Adding Semantics
• “Bottom-Up” (Classic)
Add semantic metadata to pages and databases all over the Web • Every Website becomes semantic • Everyone has to learn RDF/OWL
•
• “Top-Down” (Contemporary)
Automatically generate semantic metadata for vertical domains • Create services that provide this as an overlay to non-semantic Web • Nobody has to learn RDF/OWL
•
-- Alex Iskold
Radar Networks 15
In Practice: Hybrid Approach Works Best
Tagging
Semantic Web
Top-down
Statistics Linguistics Bottom-up
Artificial intelligence
Radar Networks 16
The Semantic Web is a Key Enabler • Moves the “intelligence” out of applications, into the data • Data becomes self-describing; Meaning of data becomes part of the data • Apps can become smarter with less work, because the data carries knowledge about what it is and how to use it • Data can be shared and linked more easily
Radar Networks 17
The Semantic Web = Open database layer for the Web
User Profiles
Web Content
Ads & Listings
Data Records
Apps & Services
Open Query Interfaces Open Data Mappings
Open Data Records
Open Rules Open Ontologies
Radar Networks 18
Semantic Web Open Standards
• RDF – Store data as “triples” • OWL – Define systems of concepts called “ontologies” • Sparql – Query data in RDF • SWRL – Define rules
• GRDDL – Transform data to RDF
Radar Networks 19
RDF “Triples”
Subject
Predicate
Object
• the subject, which is an RDF URI reference or a blank node • the predicate, which is an RDF URI reference • the object, which is an RDF URI reference, a literal or a blank node
Source: http://www.w3.org/TR/rdf-concepts/#section-triples
Radar Networks 20
Semantic Web Data is Self-Describing Linked Data
Ontologies
Definition
Definition Definition
Definition
Data Record ID
Definition
Field 1 Field 2
Value Value Value Value
Definition
Field 3 Field 4
Definition
Radar Networks 21
RDBMS vs Triplestore
Person Table ID 001 002 003 004 f_name jim nova chris lew l_name wissner spivack jones tucker
S P O
Subject Predicate
001 001 001 001 002 002 002 002 003 003 003 003 004 004 004 isA firstName lastName hasColleague isA firstName lastName hasColleague isA firstName lastName hasColleague isA firstName lastName
Object
Person Jim Wissner 002 Person Nova Spivack 003 Person Chris Jones 004 Person Lew Tucker
Colleagues Table
SRC-ID 001 001 001 001 002 002 002 002 003 003 003 003 004 004 004 004 TGT-ID 001 002 003 004 001 002 003 004 001 002 003 004 001 002 003 004
Radar Networks 22
Merging Databases in RDF is Easy
S P O S P O S P O
Radar Networks 23
The Web IS the Database!
Application A
Application B
Radar Networks 24
Are RDF/OWL the Only Way to Express Semantics?
•Other contenders:
• String
tags • Taxonomies and controlled vocabularies • Microformats • Ad hoc [name, value] pairs • Alternative semantic metadata notations
Radar Networks 25
One Semantic Web or Many?
• The answer is….Both • The Semantic Web is a web of semantic webs • Each of us may have our own semantic web…
Radar Networks 26
Why has it Taken So Long?
• The Dream of the Semantic Web has been slow to arrive • The original vision was too focused on A.I. • Technologies and tools were insufficient • Needs for open data on the Web were not strong enough • Keyword search and tagging were good enough…for a while • Lack of end-user facing killer apps • Lots of misunderstanding to clear up
Radar Networks 27
Crossing the Chasm…
• Communicating the vision
•
Focus on open data, not A.I.
• Technology progress
•
Standards & tools finally maturing
• Needs were not strong enough
Keyword search and tagging not as productive anymore • Apps need better way to share data
•
• Killer apps and content
•
Several companies are starting to expose data to the Semantic Web. Soon there will be a lot of data.
• Market Education
•
Show the market what the benefits are
Radar Networks 28
Future Outlook
• 2007 – 2009
Early-Adoption • A few killer apps emerge • Other apps start to integrate
•
• 2010 – 2020
Mainstream Adoption • Semantics widely used in Web content and apps
•
• 2020 +
Next big cycle: Reasoning and A.I. • The Intelligent Web • The Web learns and thinks collectively
•
Radar Networks 29
The Future of the Platform…
• 1980’s -- The desktop is the platform • 1990’s -- The browser is the platform • 2000’s -- The Web is the platform • 2010’s -- The Graph is the platform
• 2020’s -- The network is the platform
• 2030’s -- The body is the platform…?
Radar Networks 30
A Mainstream Application of the Semantic Web…
Radar Networks 31
What is Twine?
• Twine is a new service for managing & sharing information on the Web • Works for content, knowledge, data, or any other kinds of information • Designed for individuals and groups that need a better way to organize, search, share and keep track of their information
Radar Networks 32
How Twine Works
1. Collect or author structured or unstructured information into Twine via email, the Web or the desktop 2. Twine creates a knowledge web automatically
• • •
Understands, tags & links information automatically Automatically does further research for you on the Web Organizes information automatically
3. Provides semantic search, discovery & interest tracking 4. Helps you connect with other people & groups to grow and share knowledge webs around common interests
Radar Networks 33
Use-Cases
•Individuals
& author information about interests • Share with your friends & colleagues • Find and discover things more relevantly
• Collect
•Groups & Teams
• Manage
content & knowledge related to common interests, goals, or activities • Leverage and contribute to collective intelligence • Collaborate more productively
Radar Networks 34
Contact Info
• Visit www.twine.com to sign up for the invite beta wait-list • You can email me at nova@radarnetworks.com • My blog is at http://www.mindingtheplanet.net • Thanks!
Radar Networks 35
Rights
• This presentation is licensed under the Creative Commons Attribution License.
•
Details: This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/ or send a letter to Creative Commons, 171 Second Street, Suite 300, San Francisco, California, 94105, USA.
• If you reproduce or redistribute in whole or in part, please give attribution to Nova Spivack, with a link to http://www.mindingtheplanet.net
Radar Networks 36