SemTech 2011 features five days of presentations, panels, tutorials, announcements, new company/product launches, and conversations. It’s a place for new learning, professional networking, and business development.

www.semanticwebsummit.com

Boston , November 16-17,2010

George Roth

The Semantic WEB Summit east was a small gathering of people interested mostly in the applications of the Semantic WEB. The main trend that can be seen is that the number of real implementations is increasing and large organizations like the US Department of Defense, Best Buy, Overstock.com, BBC, Merck, etc. are more and more interested in the new technology. The main applications are in the following areas:

1)      Dealing with change of data – an alternative to data warehouses

2)      Business intelligence solutions

3)      Intelligent search

4)      Unstructured document processing

5)      Context determination from documents

This document represents a summary of the presentations with references to information sources related to those.

1. Lee Feigenbaum – VP of technology www.cambridgesemantics.com

Using EXCEL as the UI for semantic applications , data integration. They have an interesting tool, presented applications in life sciences i.e. the medical drug industry.

Presented the differences between Semantic Technology and Semantic WEB technology. The Semantic WEB Technology is a part of the Semantic Technology and refers to technologies like RDF, RDFS, SPARQL, RDFa applied to the information exposed on the Internet. The Semantic technology is a broader term and refers to Data Mining, Unstructured Data Processing, Semantic Search, heterogeneous data integration, knowledge processing, etc.

1. Michael Lang, CEO Revelytix

Revelytix is a company that works a lot with the US Military. They provide tools for:

-         collaboration in building Ontologies by communities – wiki type of system – open source can be downloaded (www.knoodle.com). We can get support from them if we want to use it. We can invite them to a meetup to show you the system remotely.

Michael presented an Ontology based architecture to build semantic applications. They are using 3 ontologies: Domain Ontology, Mapping Ontology (built using the Open Source Data mapping tool – D2RQ) , Metatdata Ontology,. He mentioned the migration from the EIW (Enterprise Information Warehouse to Enterprise Information Web). The US Department of Defense is using this model to define their future architecture.

1. Duane Degler – How to design the UI for Semantic Applications.

Sites:

www.designforcontent.com

He mentioned a very interesting project from MIT, that has a lot of Open Source widgets used to visualize large amounts of data:

http://www.simile-widgets.org

He mentioned an interesting new browser interface that can be used, named Freebase Parallax. http://www.freebase.com/labs/parallax/

Also Drupal 7 was mentioned as one of the best environment to develop Semantic Social network applications supporting: SPARQL endpoints, to create semantic metadata for the posting,

1. YY Lee – COO FirstRain – http://www.firstrain.com/

YY is the brain behind their system that data mines multiple sources of data in order to determine trends in the market.

1. Marco Neumann – Lotico – www.lotico.com

Marco is the initiator of the world Semantic Meetup trend. He collects the data from these Meetups and exposes them in a RDF Data API.

I will work with them to include the Cluj meetup in the Lotico group. Marco mentioned about a new platform that he uses. Is a news aggregator which is worthwhile to check out:

http://www.needlebase.com

1. Open Amplify – www.openamplify.com

Text Classification tool that can be used for market Sentiment analysis. The new product is named Ampliverse and is used to extract taxonomies from text.

( Is like the Open Calais platform – http://www.opencalais.com/ )

1. Making the case for Semantic Technology in Business – panel

Scott Brinker – www.chiefmartec.com – worth to read a study about marketing spending

Martin Hepp – CEO Hepp research GmbH – former STI member – he created Good Relations – the common WEB vocabulary for eCommerce used by BestBuy, Overstock.com  and O Reilly media.

http://www.heppnetz.de/projects/goodrelations/

The site presents the concept and how to use it. The Good Relations is about creating a common vocabulary to describe products including the semantic description and display aspects in order to be used by retailers who are selling certain products.

The idea is to make the product the most findable for search engines.

It can be used by pasting RDFa formats into WEB pages this making them findable by semantic search engines. I believe that this will be successful in the future because everybody’s interest (producer, marketer and consumer) is to find the product in the simplest way.

Jay Myers – www.bestbuy.com

Jay is the promoter of Semantic Technologies for www.bestbuy.com. He presented the way how the large retailer is implementing the Semantic technologies mentioning the strategic formula for this: integration with externally facing open data (tagged with RDFa) and Internal Linked Data through SPARQL endpoints used to live query internal dynamic data sources. To see the importance for best buy of analyzing this data, check out this news:

http://www.research-live.com/news/people/former-cia-man-takes-consumer-insights-role-at-best-buy/4002592.article

1. Semantic technology in the US Department of Defense

Dennis Wisnosky is the CTO and Chief Architect for the Department of Defense. The US DOD is the world larges organization having almost 2 million employees all over the world.  The main idea that Dennis is promoting is the replacement of the data Warehouse approach with Semantics for data integration. This is important for other organizations showing the the technology is maturing so that can be integrated in such a strategic  program for the US DOD. Another very interesting fact is the use of Open Source in their strategic architecture.

The common vocabulary used in this can be seen at:

http://www.commonvocabulary.army.mil/ui/group

1. Mathew Petrillo – Ontotext – Manging Unstructured Content

Ontotext is a Bulgarian company that participates in multiple European Projects and develops tools and platforms for Semantic technology.

Some of the used platforms / tools are:

-         LarKC – http://www.larkc.eu/ – the Large Knowledge Collider

-         OWLIM 3.4 – http://semanticweb.org/wiki/OWLIM – RDF management system (competes with Franz)

-         Ontotext created a JENA adapter kit for OWLLIM – to solve scalability issues for scalability

-         Ontotext uses the Joseki server for SPARQL – http://www.joseki.org/

-         Uses TopQuadrant for Ontology Management

Partners with:

-         Fluid Ops – for hosting

-         SaltLux.com – Korea

-         BPEng – http://www.bpeng.com/ – Italy

-         Profium – Finland – http://www.profium.com/ – Semantic Content Management

-         Uses GATE

1. Expert System announced a spin-off of Admantx (http://www.admantx.com/ ) a vertical build on COGITO for semantic advertising.
2. Rob Gonzales – Cambridge Semantics

Presented Anzo (http://semanticweb.com/cambridge-semantics-launches-semantic-platform-and-tools-for-non-technical-users-includes-excel-plug-in-and-web-front-end_b13239 ) a semantic platform . They applied the Semantic Platform (which uses EXCEL as UI) in applications for the bio tech companies. The applications were implemented in three areas of the companies :

a)      Assay management in drug research

b)      Manufacturing Quality Control

c)      Salesforce optimization

They announced a partnership with the CRAY supercomputer company.

1. Rachel Lovinger – Razorfish

Rachel presented the semantic applications in the media and indicated to download a very interesting document:

http://nimble.razorfish.com

They are using the Dublin Core http://dublincore.org/documents/usageguide/ for tagging, FOAF for microformats Good Relations for Common Vocabulary, RDFa. Microformat, HTML5. She mentioned how active is BBC in this space.

Technologies: Open Publish – Drupal and Open Calais);

You can see her presentation at:

http://www.slideshare.net/mediabistro/rachel-lovinger-presentation

1. Stephen Wolfram – www.WolframAlpha.com

He is an amazing guy. He explained the way how WolframAlpha works. Stephen was working for 25 years on Mathematica language implementation. He is an expert in Complexity Theory.

Here is how Wolfram Alpha works (try it out, is amazing !!!!):

1. They collected a very large amount of data from authoritative sources (not folksonomies – not from the web). The made this data computable for the Mathematica engine.
2. Every thing that is computed needs to be translated in multiple lines of Mathematica code.
3. How is the interaction with WolframAlpha ? is using natural language. For each question a disambiguation process takes place. If the system cannot determine the question in one way, a set of options is offered to the users and the user needs to pick the right option. (check out the site by addressing questions) They also use folksonomies as additional tools for disambiguation.
4. What is the answer ? They bring in everything what they know about the question: tables, graphs, maps, etc. is amazing !!!!

He used the following examples that you can try out:

“uncle’s brother son”

“gdp France”

“oil production in the world”

“banana consumption Europe”

“influenza”

“diabetes”

What is next for WolframAlpha:

-         launch a new format for documents named Computer Document Format used in creating interactive documents

-         bringing the Wolfram to the new PDAs – Droids

-         combine Mathematica with Wolfram Alpha

He showed some amazing examples. For example he took a [picture with a camera, placed it on the desktop, dragged it on a sheet with a mathematica code pasted in between brackets and put:

Edgedetect [ here was the effective picture ] – the system draw an image with the edges of the picture.

He also showed a code like:

Plot sinx — draw the sin function graph

Add Red Frame – the system put the picture in a red frame

Add yellow background – the system colored the background in yellow

He wants to create the possibility for non programmers to program in natural language. Amazing !

The papers present the experience of the Recognos Group in developing commercial semantic applications. The Recognos Group consists of three companies: Recognos Inc., San Francisco, CA, USA, Recognos Romania Ltd., Cluj-Napoca, Romania and Recognos Financial LLP, New York, USA.

The papers that were presented are:

SemTech, San Jose, CA, USA

“Semantic Technologies in the Financial Services Industry” Drew Warren, Recognos Financial, New York, USA George Roth, San Francisco, CA , Recognos Inc.
(download paper)

KEPT 2009, Cluj-Napoca, Romania

“Integrate heterogeneous data using semantic technologies”, Robert Baban, Adrian Petrescu – Recognos Romania, Cluj Napoca, Romania
(download paper)