Semantic Search Challenge Workshop Report

Thanks, once again, to all who attended IKS's Semantic Search workshop in Rome. Apologies that it has taken so long to get the content up online, we have been furiously editing video whilst trying to maintain other project commitments in the busy run-up to the holiday season.

Our main event, the Semantic Search Idol/Challenge for the most popular semantic search product, delivered some interesting results, and all of the workshop presentations and product demos can now be found below, along with some additional interviews with the key speakers and companies chosen to demonstrate.

The theme of semantic search emerged from our first IKS workshop in May 2009 as one of the key areas where the project’s participants would like more information. The objective of this second workshop was therefore to provide a clear overview of the state of the art in WORKING semantic search systems, and a better understanding of what IKS should provide to make effective semantic search possible within existing CMS systems.

Keynote Speech Peter Mika

Peter Mika, semantics expert at search giant Yahoo, kicked off the event with an introductory keynote, outlining some of the basic principles of semantic search: (apologies for the poor video quality)

The IKS Project in Focus 

After this, some of the IKS Project’s core participants gave a series of presentations detailing the key RTD themes. Wernher Behrendt the project's mentor and coordinator began with a short look at the achievements in 2009:

IKS Semantic Search Challenge

Then came the “Semantic Search Idol” challenge. The IKS Project’s organisers invited developers from eleven of the most interesting semantic search products and projects to demonstrate the best working technology currently available.

The Semantic Search challenge was kindly moderated by Stephane Croisier from Jahia, listen to what he had to say about it:

Each chosen product was given a strict 10 minute time slot in which to impress an audience composed of peers and content management industry specialists. The workshop attendees were then asked to (become semantic Simon Cowells and) vote for the best solution. The demo booklet below provides a short introduction to all the demos.

Semantic Search Demo Booklet

The links below show each demonstration, along with links to the company or project and some short interviews conducted with each group. 

And the winners are...

The results showed a clear preference for the most practical, customer-focused, working solutions: 
The overall winners were Zemanta (www.zemanata.com), a US/Slovenian-based company, whose ingenious suggestion system works as a plug-in to 12 of the most common blogging tools and small CMS solutions, including Wordpress, Typepad, Drupal and Joomla. The Zemanta system takes any fragment of plain text and automatically provides images and articles related to the topic of the text as well as relevant tags and explanatory in-text links. From the perspective of semantic search, Zemanta offers an interesting example of automatic semantic query construction by extracting key concepts from a longer piece of text.

The runners up were the team behind Semantic Mediawiki (SMW), the free extension of the wiki-system used to power Wikipedia. SMW helps to search, organise, tag, browse, evaluate, and share the wiki's content. It adds semantic annotations , which enable you to publish Semantic Web content easily, and allow the wiki to function as a collaborative database.

The bronze medal position went to SalsaDev, whose innovative search solution is in active use in the legal and financial sectors. According to the developers, SalsaDev replicates the language acquisition process of a child in order to understand highly complex, unstructured sources of data. This system outpaces the results and relevancy of keyword-limited systems and increases data consolidation, accessibility & usage.