Objective ICT-2013.4.1b High-quality machine translation

Funding Schemes: STREP, CSA

Indicative budget distribution STREP: EUR 21 million

Call: FP7-ICT-2013-10

b) High-quality machine translation

Advancing machine translation (MT) by pushing the research frontier and bridging relevant disciplines. Emphasis is placed on high-performance and easily configurable MT yielding high-quality translations suitable for publication with little or no human intervention. Expected innovations include effective hybridization of existing and emerging solutions, the ability to autonomously learn from use and human feedback, and to adapt to new situations with high portability and scalability. Work should cope with everyday language and with the need to compile translation resources dynamically from the web or enterprise repositories. Projects are expected to demonstrate the successful integration of MT within larger systems.

The call invites:

• (i) a few ambitious R&D projects investigating new approaches and research avenues well beyond the current state of the art; projects will be centred on cross-disciplinary approaches and partnerships, and address multimedia content and multimodal interaction; they will encompass everyday language as found in e.g. consumer generated content, cover multiple languages, and cater for written and/or spoken language as appropriate; technologies shall be adaptive, cope with massive volumes of content, and have a clear potential to support real-life processes;

• (ii) one broad-based support action designed to establish a unifying roadmap in each of the domains under consideration, developing a compelling research and innovation agenda until and beyond 2020, centred on close collaboration between research centres and commercial players (particularly SMEs), and based upon agreed reference architectures, common resources (standards, software, data), and shared development and evaluation facilities.