intelligent document search tools

Metaread
Rue Eugène-Marziano 15
1227 Geneve-Acacias

Auf Karte anzeigen
Firmenbeschrieb
MetaRead is a high-tech company founded in Geneva in March 2001; it is dedicated to creating intelligent document search tools. MetaRead builds on technological innovations from the University of Geneva to design and implement break-through solutions for knowledge-intensive companies. We have a unique expertise in combining artificial intelligence and linguistic analysis to come up with surprisingly efficient tools. Two lines of products were born from this particular expertise: intelligent search engines and multilingual translation assistance tools. The success of our solutions is rooted in their speed and relevance.



Schlüsselwörter
geneva meta search engine document
Marktleistungen
  • Translation-assistance tools

    a. Full-text search tools

    Most companies have to translate in several languages at least a part of their documentation (annual reports, internal documents, manuals, etc.) Year after year, they thus create a multilingual document base. Whenever a new document is to be translated, this document base should be re-used in order to avoid re-translating existing parts, as well as to find back the proper terminology and style.

    This calls for professionaltranslation-assistance tools. Finding back an adequate translation requires not only to produce a list of documents containing given keywords; it requires a tool which allows to display the source and target documents side by side, with an automatic alignment of the versions. It also requires a tool which also allows to display dozens of reference documents one after the other, in order to find the proper context and terminology. Last but not least, it requires a tool which supports any written language and which is web-compatible (in order to allow external translators to use it as well).

    MetaRead has created two different tools for this purpose: ISIView and ISIRef were designed and implemented by a team of professional translators for their own needs. These tools are currently used in several international organizations. They can be tested online on this website.

    b. Terminology management

    Any company with a highly-advanced expertise knows how crucial it is to manage properly its internal terminology. Each proprietary concept is a competitive advantage; thus each concept must be clearly named and defined. This is particularly critical to ensure that employees worldwide, including new comers, have the same understanding of concepts and use the same vocabulary.


  • Search engines

    a. Text categorization Many companies manage thousands or millions of documents without a fully-fledged document management system. At some point they need at least to organize their documents in various categories or types. Defining those categories is a difficult task, but when this is done an even more difficult task must be performed: distributing several gigabytes of documents over those categories. MetaRead has several solutions to automate the distribution of millions of documents across dozens or hundreds of categories. This is particularly helpful not only to manage corporate documents, but also to classify technical documentation, organize a knowledge base, or create intelligent catalogues.

    b. Intelligent search tools

    Searching an intranet content with common search engines (such as Google or Altavista), or with proprietary engines under a document management software, is usually either inefficient or slow. When it is inefficient, it retrieves lots of documents which are completely irrelevant to the search context; for the user, this results in a loss of time. When it is slow, it also results in a loss of time - and sometimes the result may be both slow and inefficient! MetaRead has developed a unique search engine which produces amazingly relevant results in a time which is usually counted in milliseconds. We combine the use of neural networks with customized semantic networks to select semantically-heavy keywords and to predict the most relevant categories of documents. Whatever the search engine that may be used, we are able to boost its efficiency both by limiting its search field and by orienting it towards the most relevant types of documents. This method mathematically improves the results of any search operation!


Eintrag erstellt: 26.08.2004, Letzte Änderung: 26.08.2004