Notes from Wednesday afternoon Semantic Technology Conference

Here are my notes from Wednesday afternoon/evening at www.semantic-conference.com

==============================================================================
Wednesday afternoon, 6/17/2009

------------------------------------------------------------------------------
12:30PM - 1:30PM

Wolfram|Alpha - An Interview
Foltz-Smith, Russell, Wolfram|Alpha - www.wolframalpha.com
Spivack, Nova, Radar Networks (Twine.com)
Business / Non-Technical
Semantic Query

Systematic (computable) knowledge (numerical, emphirical).
e.g.,: Physics, finance info, weather data;
(as opposed to Non-computable: e.g., 
 latest opinion on who Britney Spears is dating)
Calculate viz enumerate
Does not memoize results.
Syntax: hard to make understandable input queries
No top-level ontology
Engine is Mathematica
Business models:
- API pay-per-use
- subscription
- advertising
Can be used in a limited manner as cross-domain discovery
Picks interesting and tractable topics
Want to have coherent way to let users access knowledge model
- perhaps might use SW technology
Education
- learn how to compute viz how to orchestra chain on calculations
Future: accessible data curration tools so user can
  and data so it can scale

------------------------------------------------------------------------------
Conference Sessions 2:30PM - 3:30PM

--------------------------------------------------
Next Generation Protege: OWL 2 Ontology Development for Web 2.0
Tudorache, Tania, Stanford University
Redmond, Timothy, Stanford University
Noy, Natasha, Stanford University
Technology - Intermediate
Knowledge Engineering and Management

Ontology engineering (like TopBraid composer, but OpenSource)

OWL 2.0 via Protege 4.0 (can handle OWL 1.1)
WebProtege
- protege.stanford.edu
- Instant access to ontologies
- Online collaborative

------------------------------------------------------------------------------
Hybrid Approaches to Taxonomy & Folksonomy
Wlodarczyk, Paul, Earley & Associates
Beatch, Richard, Earley and Associates
Business / Non-Technical
Taxonomies

Co-existence
- used side-by-side but separate
- EX: library of congress
- Variation: tag mediation
  Vetting and editing of tags
  Pros: weed out inappropriate, misspellings, plural
  Cons: effort, perceived lack of trust of users, who knows better

Tag-influenced taxonomy
- tags input as candidate terms to enrich taxonomy
  Find new terminology/concepts (synonyms, popular language)
- Requires formal vetting process

Taxonomy influenced tagging
- present choices/suggestions to user from controlled set of terms/tags
  drop down; check boxes; type ahead, tree view
- option to enter new tag
- enforces consistency
- offers structure
- EX: ZigTag (social bookmarking site)
- EX: TextWise
- EX: Buzzillions.com
PROS: more consistency; findability; relationships leveraged;
      realistic for enterprise

Tag hierarchies
- User powered
  social approach
  bogus possible
  can be esoteric
  small population of contributors
  EX: LibraryThing
  EX: Flickr - machine readable tags
  EX: moat-project.org : adding "meaning" to triplet user/resource/tag
- Automatic derivation  
  Statistical/clustering algorithms
  heymann.stanford.edu/taghierarchy.html
  EX: clustering at flickr
  Not mature; time-sensitve; community-sensitive; ambiguous, 
     improves with volume

richard@earley.com
paul@earley.com

web: www.early.com
blog: sethearly.wordpress.com
twitter: earlytaxonomy

------------------------------------------------------------------------------
Conference Sessions 3:45PM - 4:45PM

Semantics at Google: RDFa, Microformats, and more
Hansson, Othar, Google, Inc.
Goel, Kavi, Google
Technology - Introductory
Publishing

Rich Snippet
- Text between title and URL in Google serach results
- Sites encouraged to use RDFa or microformats
- EX: vcard; hreview; 

Current (out of a million head pages):
- 11,000 vcard
- 5600 foaf
- 2500 hReview
- hProduct

Open
- RDFa and microformats
- No back-channels to Google
- Engage with community

------------------------------------------------------------------------------
EXHIBIT

neon-toolkit.org - ontology engineering environment; Open Source
Syntactica.com - Entity Extraction (like Calais but code, not a service);
                 (Open Source ??)
Anzo:  software vendor; 
       Open Source http://www.openanzo.org/
       excel (from distributed source) 
         -> RDF
           -> anzo (distributed) server(s)
             -> faceted browsing
Metatomix: SW platform - used in justice, finance, ...

------------------------------------------------------------------------------
Conference Sessions 5:00PM - 6:00PM

Freemix - Social networking for data.
          Wouldn't it be nice if your data had friends, too?
Miller, Eric, Zepheira
Wood, David, Zepheira
Technology - Introductory
Data Integration and Mashups

Zepheira
- Professional Services Organization
- Freemix - a tool

Users want THEIR data THEIR way
Data is there, but it is locked up in application
Desire to take that data from one app and use in another or to drive
 other views

Freemix
- rapidly combine private and public information, style it, share it
- Gain new insights into patterns in data
- Built on open source, open standards
  Simile, OCLC, Aduna, Topaz, Fedora Commons, Cleveland Clinic, Akara

Data A                  View x
Data B ---> Freemix --> Viex y
...                     ...

Freemix
- Easy to expose collection via web
- Connect people around info
  Data from one access point
  Access and connect data across multiple collections of data
- Creating and preserving community
  Share best practices
  Establish network of trust and ...

How is it done
- Empower users
  Create your own view
- Build a community around your data
  Help each other curate and connect as needed
- Skip the supermodel
  Leave data in the systems, wrap and expose using web as platform

Freemix platform viz Freemix.it
- .it: on web - your data is public
- platform: can keep your data private

Freemix
- An easy-to-use version of Simile Exhibit:
  http://www.simile-widgets.org/exhibit/
- But more than just making exhibits
- Can get underlying data out as RDF (serialized in JSON)

Coming soon
- mixing and integrate local and public web content
- human annotation and data augmentation
- scalability, usability, accessibility, I18N
- Freemix Enterprise

em@zepheira.com
david@zepheira.com

Need
- open source
- open standards
- open content
- open content licensing

--------------------------------------------------
Introducing and Applying Semantic Technologies to the Financial Services Industry
Warren, Drew, Recognos Financial LLC
Roth, George, Recognos Inc.
Business / Non-Technical
Industry: Financial

Books and other info sources 
- Semantic WEB Programming 
  ISBN 978‐0‐470‐41801‐7 
- Semantic WEB for the Working Ontologist  
  ISBN 978‐0‐12‐373556‐0 
- The Text Mining Hand Book
  ISBN 978‐0‐521‐83657‐9

------------------------------------------------------------------------------
Evening Events 6:00PM - 8:00PM

Linked Data Gathering
Idehen, Kingsley, OpenLink Software, Inc.
Technology - Introductory
Linked Data

Talis; Squin; ClaraBridge; 
Linked Data coming to Britain and US governments (via Tim Berners Lee)
  see whitehouse.gov/innovate or whitehouse.gov/blog