The NEMO (laNguagE Modules develOpment) project, whose scientific and technological development is led by Zanasi & Partners, has been funded by the European Commission (EC) within the European Defence Fund (EDF) between 2024 and 2029 with Grant Agreement no. 101167770.
The NEMO project aims to develop a demonstrator directly integrated into the activities of the military end user. The demonstrator will be a software prototype offering a complete range of different Human Language Technology (HLT) functionalities, with an interactive and user-friendly interface for user-driven adaptation and with special attention to AI explainability and performance evaluation.
The project will be executed by a consortium of 15 partners from 9 European countries, including research institutes and companies at the forefront of HLTs as well as the Bulgarian and the Romanian Ministries of Defence.
The software will analyse information encoded in texts, audio messages, and images expressed not only in all EU languages and other European languages such as Norwegian, Russian, and Ukrainian, but also in Arabic, Mandarin Chinese, Japanese, Farsi, and Hindi. As a result, the software will spot keywords, classify text, summarise key ideas, and more.
To this end, an innovative approach based on three pillars will be applied.
First, the most advanced algorithms and expertise developed by industrial and academic institutions at the forefront of HLTs will be used as starting points for NEMO.
Second, the developed tools will be adapted to the Defence domain, characterised by its vocabulary and semantics. Hence, the partners in the defence sector will play a crucial role in the project proposal.
Lastly, the proposal must enable users to perform data annotation, correction and validation in accordance with the requirements defined in the call. To this end, an approach based on Transformer Models and Large Language Models will be developed.
Documentation
The NEMO Consortium
