Additionally we have added new features such as support for multiple Linked Open Data cloud sources or explanation functionality for concept ranking that essentially exposes the particulars of the concept similarity algorithms to the interested user. Other technical improvements worth mentioning are the handling of multiple concepts in one query and the so called auto populate-feature. The latter enables semi-automatic addition of new concepts e.g., to a thesaurus.

An evaluation of the Response Alignment Service based on two data sets helped us assess the overall correctness of the implemented Response Alignment Service algorithms, and additionally provide insights into possible improvements of the concept ranking and the lookup process.

Use Case Semantic Desktop (SD)
In Refinder these services are exploited for two main purposes: one, to turn information from external sources into shareable social objects and make them available for a team; and two, to enrich existing Refinder items (like documents, web links, persons, or topics) with contextually relevant descriptions that form an information graph around those items. These descriptions are considered during search and recommendations calculation, and directly affect Refinder’s ranking mechanism. So, users are supported in quickly and effectively accessing the information they need to get their work done.

Currently, the LASSO add-on for Refinder is available in a private alpha mode. Contact us at research@gnowsis.com to take a look at it.

Use Case Enterprise Collaboration (EC)
In the PoolParty Thesaurus Management System this LASSO service is used to match concept labels from a SKOS thesaurus against labels from LOD resources in order to find potential alignment candidates.
Labels of related, narrower and broader concepts are used as contextual information and compared to literals (labels, abstracts, category labels) which are associated with the LOD alignment candidates. With this additional information the disambiguation part of entity resolution is improved.

PoolParty uses the auto-populate feature to help thesaurus managers to extend their models semi-automatically with resources from Linked Data sources that are similar to existing concepts. Here the system tries to find DBpedia categories for the existing concepts and then suggests instances of these categories which thesaurus managers can add to their thesauri.

Use Case Inspiration Services (IS)
In the Neurovation idea platform the LASSO lookup service is used to service the user with valueable inspirations (word-image combinations) during the idea generation process. The user first defines an idea title which is interpreted as the input context. The context is sent to the LASSO service which identifies related concepts in the LOD cloud matching the given context. The LASSO ranking service puts the matching concepts in order. The user can browse through the results by “broadening” or “narrowing” the view on the inspirations.

Handling multiple concepts and LOD sources are utilized by the inspirations service to offer the user inspirational and diverse word-image combinations of a wide range of topics. The user gets inspired by “aha-effects” that occur from the received inspirations.

Within the i-Praxis track, the LASSO consortium presented the LASSO approach and the three involved use cases in Session
A LASSO for Linked Data – 3 Use Cases to improve personal knowledge management, document management and creativity techniques by leveraging Linked Data.

Link to the slides on Slideshare.

Correspondingly to a local concept coming from the PoolParty thesaurus management application present in the Use Case Enterprise Collaboration (EC), a lookup to the Linked Open Data (LOD) cloud is carried out. The results from the LOD cloud are then ranked according to various algorithms, depending on Use Case and subsequent processing.

The second prototype scheduled for beginning of 2012 will include full integration into the LASSO framework, refined ranking algorithms and improved Linked Open Data cloud support.

• Knowledge Management (WissensCamp)
• Design (DesignCamp)
• Politics (PolitCamp)
• Mobile Development (iCamp)

The LASSO project was introduced by the LASSO Team at WissensCamp. The german slides are available on the web.

One of the three areas we are working on is the “Enterprise Use Case”, where we eventually want to show that a knowledge model (like a taxnomy or a thesaurus) is a better source of tag and content recommendations when it is aligned with resources from Linked Open Data sources. In order to cost-efficiently create such LOD-aligned thesauri we first focus on algorithms that help thesaurus creators to establish such alignments with thesaurus management tools like PoolParty.

We currently work on services that (semi-)automatically look up and match concepts from a locally stored knowledge model with resources from the LOD cloud. Once a couple of local concepts are matched and anchored with LOD resources, we want to leverage these links in order to semi-automatically populate parts of the local knowledge model with fitting concepts.

We want to demonstrate that the more connections between concepts of the local knowledge model are established with the LOD cloud, the better further concept disambiguation and auto-population becomes. This idea is reflected in the LASSO metaphor: The LASSO is used to catch fitting concepts from the LOD cloud – and the more connections get established, the tighter the grip of the LASSO becomes.

LASSO is a two-year research project funded by FIT-IT Semantic Systems and Services, a program of the Autrian Research Promotion Agency (FFG).

On Mai 4th, 2010 the official Kick-Off meeting was held in Vienna, Austria. 8 participants from 4 research and technology partners came together to kick start the project and showed their engagement in interestring discussions during the plenary session.