Infom@gic

Janvier 2008

Press release

Pertimm develops a new generation web search engine as key component of the Infom@gic UIMA project.

Paris, January 21, 2008 – Pertimm is a French publisher of "search engine" solutions and a key player in "Infom@gic", a research and development project in the Cap Digital Competitiveness Cluster, with a focus on knowledge engineering technologies.

Pertimm, together with more than 25 private and public sector partners, is providing its expertise in terms of semantic relevance and information searches.

"Intelligent" information searching

The Web search engines available to the public today are based on an analysis of keywords or characters, and are based on a quantitative approach to the data (this latter being classed on the basis of popularity – the page rank – and the date of creation). Despite their extensive use, internet users are aware that these tools are limited: the need to express the subject of the search precisely and concisely, the existence of silences (relevant pages for a given search but not returned by the engine or not indexed), lack of relevance of the responses, etc.

As part of the "Infom@gic" project Pertimm has developed a new generation web 3.0 search engine making use of original functionalities. The French publisher aims to demonstrate the utility and the added value of a linguistic and semantic based engine for information searches. Intended for use by all users the new tool will provide an effective alternative to today’s leading Internet search engine providers, in particular through the application of its "Pertimmiseurs" technology which has already been successfully implemented for more than five years on its clients’ Intranets.

Original search functionalities

Unlike current engines the Pertimm tool applies a semantic processing with Pertimmiseurs. The taking into account of the linguistic, the ontological, named entities and concepts, as well as navigation through the results by co-occurrences, means that it can offer users a much wider and more relevant spread of responses than with normal engines. This is all the more a reality as, unlike existing engines, it offers the user the opportunity to carry out a search applying an extended corpus of words. At the moment, beyond 3 or 4 words, this leads to a failure or non-relevant results. With Pertimm, the more words input by the user, the greater the relevancy of the search results.

Another innovation is the inclusion of the qualitative aspect for the sources rather than the Page Rank. This enables internet users to access primary articles even if they were created some considerable time ago or are only infrequently viewed.

This new search engine is thus particularly suitable for highly specific searches or when a user does not know exactly how to state what he/she is looking for.

Pertimm, a major player in the "Infom@gic" project

Drawing on its expertise in terms of multimedia search engines, Pertimm has a central and unifying role in the areas of the “Infom@gic” Project relating to searches:

• Management in the context of the SP1 "Searching for electronic information" sub-project for all aspects relating to information searches, from its setting up and the organization of the project right up to the demonstration of its feasibility.
• Coordination of the use of the “Web search” application in the context of the UIMA (Unstructured Information Management Architecture) standard, with the partners in the SP2 "Information extraction" sub-projects.
• Monitoring of the work in the image field and the provision of its expertise concerning specific points.
• Involvement in work relating to the corpus of the patents of France Innovation Scientifique et Transfert, the CNRS technology marketing structure.
• Involvement in the SP5 sub-project, coordinated by INA, for accessing digital assets.

About Pertimm

Founded in 1997 by three engineers, Pertimm is a French software publisher with a specialization in multimedia search engines, using linguistics and semantics to enable the “discovery of knowledge” and this regardless of the nature, the structure or the organization of the data media.

Located in France, Asnières sur Seine (92) and Valence (26), and in the United States, Pertimm has a staff of 20. The company was awarded the Anvar Innovative Start-up Label prize in 2004. A number of companies choose Pertimm technology in 2006 and 2007, including Meetic, Neuf Cegetel, Pages Jaunes, CNAMTS, etc.

About Infom@gic

Coordinated by Thalès, and bringing together more than 25 private (EADS, Xerox, Temis, etc.) and public sector partners (CEA, INA, French University Research Laboratories, such as LIP6, etc.), the “Infom@gic” project, for which development work started in January 2006, aims to set up an industrial laboratory for the selection, testing, integration and validation of applications incorporating the best French technologies.

This is based around a shared interoperable platform covering the 3 fundamental technological components in the field: advanced search engines, knowledge extraction and the merging of multimedia information, and covering all existing source types: structured data, text, images & video and sound. "Infom@gic" by definition will cover an extremely varied range of applications such as digital asset management, marketing challenges, e-Education, research and business intelligence…

Aimed at teachers, corporate management, financial analysts, the military and researchers, this cooperative project is being supported as a part of the Image Cap Digital Competitiveness cluster.

By creating a unifying dynamic among the players involved, "Infom@gic" will facilitate the development of innovative and world-class, interoperable and easily applicable technical solutions covering a very wide range of contexts.

Pertimm press contact

Silvy Bikok :
Tél. : +33(0)1 47 33 88 49

Infom@gic press contact

Pierre Hoogstoel :
Pierre.hoogstoel@fr.thalesgroup.com