CNGL Demonstrators

Demonstrator systems are a core part of CNGL research. The demonstrators provide focal points for project cohesion and collaboration, combining technologies and teams from across CNGL.

The demonstrators are essential for cross track and overall project evaluation and provide platforms for research and experimentation across the CNGL. They showcase CNGL technologies to the outside world. CNGL demonstrator systems instantiate core use scenarios in the space defined by the localisation cube:

The Bulk Localisation Workflow (BLW) scenario targets large volume localisation tasks with and without human pre- and post-editing familiar from current large localisation projects. The focus is on predictable corporate content, automation (MT) and the optimal integration of crowdsourcing (where applicable) in an off-line localisation process with modest levels of personalisation and standard print and full screen based access modalities. The Personalised Multilingual Customer Care (PMCC) scenario focuses on users interacting with on-line and perishable digital corporate and user-based (product blogs) content, providing for frequent updates, speech-based interaction modalities (in addition to the more traditional modalities) and sophisticated levels of personalisation in real time interactions, without human pre- and post-processing interventions. Finally, the Personalised Multilingual Social Networking (PMSN) scenario focuses on user generated (in contrast to corporate) and highly perishable content prevalent on social networking sites, with high levels of personalisation and full use of all access modalities, to put networking sites in contact across linguistic barriers. CNGL demonstrator systems evolve according to a five year research programme with base-lines evolving into more sophisticated systems in subsequent project years:

Years four and five will see the emergence of an adaptive Unified Localisation Factory (ULF), which can instantiate any point in the localisation cube on demand with configurable quality and speed. In order to realise this trajectory, workflows need to be highly flexible, reconfigurable and adaptive: all demonstrator systems (starting with year one) are based on the shared CNGL components framework (including components from the CNGL academic and industry partners) and web services based architectures, where components cluster into sub-scenarios, and sub-scenarios provide important parts of the CNGL demonstrator systems.