Many industries are coping with the challenge of working with a growing amount of digital content across multiple domains and languages. The "Open Framework for e-services for multilingual and semantic enrichment" project (FREME) is designed to help by providing a commercial-grade framework of e-services that will improve existing digital content management techniques.
FREME enrichment services will provide a reusable set of open Application Program Interfaces and Graphical User Interfaces and address the whole content value chain: content creation (or authoring), content translation/localization, publishing and access to content including cross-language sharing and personalized content recommendations.
A two year project funded by the European Commission under the Horizon2020 research program, FREME started in February 2015 and is intended to open new opportunities for all sectors involved in digital content management and strengthen the position of the European Union as provider of innovative multilingual products and services based on digital content and data.
The FREME Project intends to provide a set of interfaces for:
* The enrichment of textual digital content by adding linguistic information in a language or languages other than the language of the content itself
* And semantic enrichment, which uses metadata, Natural Language Processing Interchange Format (NIF), semantic richness provided by Linked Open Vocabularies and large and prominent repositories in the Linked Open Data (for example, DBpedia, Freebase, and others) to transform unstructured content into its structured representation.
To bring FREME to market, FREME e-Services will address:
- Authoring and publishing multilingually and semantically enriched eBooks
- Integrating semantic enrichment into multilingual content in translation and localisation
- Enhancing the cross-language sharing and access to open agricultural and food data
- FREME-empowered personalized content recommendations.
After the first six months of Project, FREME provides four e-Services. The e-Services provided by FREME will be capable to process (harvest and analyze) content, capture datasets, and add value throughout content and data value chains across sectors, countries, and languages. They are:
* e-Entity. based on entity recognition software and existing linked entity datasets from Linked Data and Open Data
* e-Link. based on the Natural Language Processing Interchange Format (NIF) and RDF;
* e-Publishing. based on cloud content authoring environment (for example e-books, technical documentation, marketing materials etc.) and its export for publishing in digital publication (EPUB3) format ng, and
* e-Translation. based on cloud machine translation services for building custom machine translation systems.
Each one can be accessed as a Web service via a RESTFul interface. The REST API design has been chosen to permit an easy integration of FREME tools written in various programming languages. All FREME e-Services use the Natural Language Processing Interchange Format, NIF, as the broker format for access to the services and for interaction between the services themselves.
Consider this sample use case involving accessing agricultural and food data. Today users often cannot grasp the content in their native language. Search is possible only in the language of the indexed documents. There is low quality of translations, and terminology is not correctly translated. Content metadata is of low quality and incomplete, prohibiting good discovery services, and content is not connected to external data sources.
FREME will improve this situation in several ways: e-Translation will allow users to access content in their own language. e-Entity will be used to detect and search relevant entities in user queries; in combination with e-Translation, this will allow users to discover content across language barriers and explore structured data with their native language. e-Terminology will help to create high quality terminology efficiently. e-Link will make it possible to connect the terminology with external data sources. e-Publishing and e-Internationalisation will provide standardised representation of content including enrichment information. This will be the basis for further content re-use.
“Since FREME objectives span a wide range of business and research communities, we have discussed FREME at many events, ranging from cutting edge research conferences or industry specific events to EU networking, related to the communities of data, language technology and public sector information,” says project coordinator, Professor Felix Sasaki from DFKI, one of the largest non-profit contract research institutes in the field of innovative software technology based on Artificial Intelligence (AI) methods.