Community Research and Development Information Service - CORDIS


FREME Report Summary

Project ID: 644771
Funded under: H2020-EU.

Periodic Reporting for period 1 - FREME (Open Framework of E-Services for Multilingual and Semantic Enrichment of Digital Content)

Reporting period: 2015-02-01 to 2016-01-31

Summary of the context and overall objectives of the project

The aim of FREME is to build a framework for multilingual and semantic enrichment of digital content. The framework provides a set of set of application programming interfaces (APIs) and graphical user interfaces (GUIs) to access six e-Services:
1) e-Internationalisation based on Internationalisation Tag Set (ITS) Version 2.0;
2) e-Link based on linked data and Natural Language Processing Interchange Format (NIF);
3) e-Entity based on named entity recognition software and existing or user provided linked datasets;
4) e-Terminology based on cloud terminology services for terminology management and terminology annotation web service;
5) e-Translation based on cloud machine translation services for building custom machine translation systems; and
6) e-Publishing based on cloud content authoring environment and its export for publishing in Electronic Publication (EPUB3) format.

The FREME framework offers the following benefits to the above e-services and also potentially to others:
* Growing support of widely used digital content formats as input and output for enrichment;
* Re-usability of the framework in various application contexts, driven by the FREME business cases;
* A set of data sets available on-the-fly for multilingual and semantic enrichment;
* Adaptability of processing to domains and user provided data sets;
* Deployment of key standards for language and data processing; and
* Achieving interoperability between data and language processing and easing extensibility with e-services by putting standards in action.

The FREME framework is built triggered by four different business cases:
* BC1 “Authoring and publishing of multilingually and semantically enriched eBooks”
* BC2 “Integrating semantic enrichment into multilingual content in translation and localisation”
* BC3 “Enhancing the cross-language sharing and access to open agricultural and food data”
* BC4 “FREME-empowered personalised Web content recommendations”

These four business cases are real life examples for FREME to validate that multilingual and semantic technologies are ready for integration in an innovative way.
FREME contributes to innovation in business models and growth of businesses: FREME supports its technological development with four business plans that accompany the four business cases in their journey “from the lab to the market”. FREME also leads to the extension of job profiles and lays the basis for the acquisition of new skills in the realm of digital content management.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During the first year, project objectives and milestones were achieved, project deliverables were submitted and, as a whole, the project work was adherent to the work plan in relation to the Annex 1 to the Grant Agreement (Description of Action, DoA). Overall, the project made excellent progress (the project fully achieved its objectives and technical goals for the first year and even exceeded expectations). The work done in each work package (WP) is described below.
The project started with an extensive phase of requirements gathering for FREME and specification of FREME = Tasks 1.1, 2.1, 3.1, 4.1, 5.1. The FREME business case partners iMinds (BC 1), Vistatec (BC2), Agroknow (BC3) and Wripl (BC4) formulated their requirements with regards to: language support, domains, format coverage and data sets to be used for enrichment. Details are provided in related deliverables D2.1, 3.1, 4.1 and 5.1. The discussion with the technology partners DFKI, InfAI and Tilde took place on the FREME mailing list, the FREME GitHub repository, a dedicated requirements gathering online form, regular phone calls and two face-to-face meetings. Results of the requirements gatherings are summarized in D1.1.
From the beginning FREME was not conceived as a platform but as a framework of components, with different types of users:
0) Technology specialist in digital content management, Big Data, multilingual and semantic technologies
1) Data expert and technology user
2) Interface developer
3) End user (content creator, translator / localiser, publisher, and other users that this or that way create, process, or consume content)
As an implementation of the FREME specific objective “Objective 5: Extension of job profiles and facilitating acquisition of new skills”, the requirements gathering led to an initial list of existing job profiles who could benefit from FREME in the four business cases. Also, the four business cases identified several new job profiles that could benefit from FREME.
The initial specification of FREME focused on providing APIs. FREME is being developed as a web server based architecture. Software components by business case partners are using the FREME API. Services external to FREME are for example e-Translation and e-Terminology; they exist outside the FREME framework. Services which are available as part of the framework are for example e-Link or e-Publishing. The conversion services allow FREME to handle a variety of digital content formats and rely on e-Internationalization. Pipelines allow to chain several e-Services.
The architecture shows a key aspect of FREME: it is not a monolithic platform but a set of components. The broker is a mandatory component to allow access to other components.
FREME was presented in various occasions also to get feedback from outside the project consortium. The requirements analysis was updated also with feedback from the BC partners in M10, see a summary in D1.3 and details in the BC specific deliverables D2.2, D3.2, D4.2 and D5.2.
The updated requirements revealed that the BC partners continued to deploy the framework in the same way as they did for previous releases. To ease modularization, the software architecture was updated and put in context of other initiatives and projects like Big Data Europe, Apache Stanbol or S4.
Besides the forehand mentioned modularization and component / building block approach, a key feature of FREME is that it provides semantic and multilingual processing based on standards. This improves interoperability and eases re-use or re-combination of e-Services. The Task 1.2 Data Innovation and Take Up laid the foundation for this approach, by putting the following standards in action:
* Internationalization Tag Set (ITS) 2.0;
* Natural Language Processing Interchange Format (NIF) 2.0;
* The OntoLex Lemon model for representing lexical and ontology information; and
* The Linked Data technology stack, which encompasses URIs, RDF and SPARQL.
Via the Task 1.3 Technology Transfer we implemented the specific FREME objective “Objective 2: Integration and transfer of technologies”. In addition to software components for processing ITS 2.0 and NIF 2.0, the following technologies were made available to the business case partners:
* FREME Named Entity Recognition (NER) as the basis for e-Entity;
* Tilde Terminology as the basis for e-Terminology;
* Tilde MT as the basis for e-Translation; and
* iMinds cloud authoring environment as the basis for e-Publishing.
Technology transfer in FREME is based on an agile software development approach. Milestones for the minor releases of the framework (0.1 – 0.5) are complemented by: continuous bug reporting and bug fixing; discussion and implementation of new features; and intensive testing phases to assure the robustness of the framework.
The Task 1.4 Planning of Alpha and Beta piloting and validation documented via D1.2 plans of BC partners for deploying the FREME framework and e-Services. The actual implementation took place in Task 1.5 First pilot of the FREME framework and e-Services – Alpha version, and in the BC specific counterparts Tasks 2.2, 3.2, 4.2 and 5.2. FREME Alpha is documented in D1.4 and its implementation in BCs in D2.3, 3.3., 4.3 and 5.3. Table 2 summarizes key aspects of the Alpha pilot with regards to the four business cases. This task was key for implementing the specific FREME objective “Objective 1: Innovation in content and data value chains” and the objective “Objective 3: Validation within real world business cases”. Alpha validation in several aspects went beyond what was planned in the DoA.
The framework has been developed including business modelling and planning, via the related Tasks 2.4, 3.4, 4.4 and 5.4 on Market validation of FREME business cases - business model and business plan. ISMB led the tasks which implements the specific objective “Objective 4: Innovation in business models and growth of businesses”. ISMB provided the methodological basis, as well the tooling. Using the Lean methodology and Osterwalder’s business model canvas, several one-to-one calls with the business case partners were conducted. Like the software development methodology, the business modelling is following an agile approach. The intermediate outcomes can be summarized as follows.
For BC1: iMinds sees commercial opportunities for the development of a digital learning platform by a digital publisher operating in the Flemish area. This takes the initial exploitation plan from the DoA into account with a focus of a sub segment of the electronic publishing industry.
For BC2: Vistatec sees added value for localisation processes by integrating FREME e-Services. New features based on semantic and multilingual enrichment will allow the localization company to target additional sources of revenue and expand their offerings. For Vistatec it is important to add extra-value for the same price in this highly competitive industry.
For BC3: Agroknow is offering FREME enabled services to various types of customers: large international organisations, small International organisations that work in specific thematic areas, small publishers that want to promote their journals via a service with high visibility, and publishers to which Agroknow is providing the metadata records for agricultural research information. In all these areas, FREME based enrichment provides new value to Agroknow and their customers.
For BC4: Wripl sees added value in data analytics and added content focused services such as search engine optimization and data visualisation. New features were introduced due to FREME related services. The main feature introduced relates to data analytics and visualisation. This enables Wripl customers to analyse the users engagement related to entities and terminology. The goal is to support users in making decisions related to content strategies, e.g.: what content topic is needed to be able to increase the number of publications. Currently Wripl is seeking trial partners for multilingual enrichment, in German, French and Spanish.
In order to engage with groups outside the project consortium, a dissemination and impact maximisation plan (see Task 6.1 and D6.1) was created. The outcome of dissemination activities was documented publicly in M12 via D6.2.
FREME installed a project website and social media channels, produced dissemination material and academic dissemination via papers, and participated in workshops or other events for selected target groups, with a focus on the data value chain. The branding and web visibility (see Task 6.2) was presented in a homogenous visual identity in all public channels and dedicated materials (press releases, newsletters, flyers etc.), with a specific effort for European Data Forum 2015.
The Task 6.3 on engagement with the data value chain industry was highly successful. Already more than 30 companies have expressed interest in at least one of the FREME assets. The details include small versus big companies, interest in horizontal versus vertical FREME assets, and interest by all four FREME user types: business cases’ end users (user type 3), third parties interested in using FREME open source framework (type 2), FREME e-Services prospects (type 1), and technological and/or commercial partners (type 0). Via Task 6.4 we were also successful in attracting certain professional communities, including the big data area, standardisation, language technology, cultural heritage and selected industry sectors. Technology groups that showed high interest in FREME include XML developers, translation and localisation, and the linked data and public data communities. Especially the latter groups (but not only them) were targeted in liaisons to various EU projects, contributing to the dissemination goals of: growth of visibility, experimentation with FREME technological assets, organizational support and growth of the pool of FREME prospects.
FREME also disseminated scientific results (Task 6.5) of year one to the scientific and various industry related communities. Details about these efforts are report in D6.2, see especially table 17 of that report. We assume that this effort will also contribute to the sustainability of FREME and build the pathway for further journeys from the lab to the market via FREME, taking current research results into account.
With regards IPR, exploitation and sustainability (Task 6.6), the consortium has explored various sustainability strategies but not yet committed to a single strategy. A key aspect of pushing one or several strategies forward is the availability of the FREME framework as a set of open source components with suitable licenses to ease commercialisation, with a clear distinction from external e-Services. This setup should be a basis for both individual as well as joint exploitation plans after the end of the project.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

After its first year, FREME has made significant progress on all four areas of expected specific impact.
1) Innovation and technology transfer in multilingual data harvesting and analytics solutions and services: The FREME framework has been established in an alpha version. The six e-Services have been developed in close interaction with the business case partners. Technologies from the realm of language (e.g. ITS 2.0) and data (e.g. NIF, Lemon OntoLex and the general linked data stack) have been successfully transferred into the business case partners applications. Specific care has been taken of making linked data sources useful for people who are not (linked) data experts, e.g. by mechanisms in the e-Link service.
2) Creation of data products and services that the action participants will bring to market in a sustainable fashion: all business case partners and technology partners have cooperated closely during the development of FREME alpha. The validation of FREME alpha has been undertaken with an agile approach, with the technology partners in the project taking business case partners feedback constantly into account. Of specific value are the business case partners looking into the competitive market of enrichment solutions for digital content. This helped to shape the FREME framework as a set of building blocks rather than a single platform, enabling extension with 3rd party services easily by providing standardized data formats and interfaces.
3) Collection of data in one domain, usage in a different one: the usage of FREME across services has demonstrated the feasibility of data re-use across domains. An example is the application of the Agrovoc thesaurus. This resource has been developed in the agricultural domain. In FREME it is being used via e-Terminology for multilingual and cross lingual enrichment. We expect in the second year of the project and for the long-term sustainability much more examples of cross- domain data re-use and data-integration.
4) Adhere to business needs in terms of data quality management, cost of data curation and technical needs on all levels (from character encoding to multilingual thesauri + ontologies): The FREME consequently has been developed with deployment of standardised technologies, as already laid out in the DoA: EPUB, HTML5, ITS, Linked Data technologies / NIF, XLIFF. FREME alpha consists of open source components that are tested continuously and that are developed in an open software repository. The software development of the FREME already has created interest from outside the consortium. We assume that this will help to foster new business cases and the integration of new technologies in the framework.
FREME made significant progress in all areas of general socioeconomic impact that are targeted.
At the end of year 1, already more than 30 companies have shown interest in using FREME, with various types of users involved. Several connections have been made to other players in the European Data landscape, including a rich set of liaisons to other EU-funded R&I projects. The FREME project has been highly visible in social media at large.

Related information

Record Number: 186657 / Last updated on: 2016-07-14
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top