arma et scientia ex officio

Call results by calendar year

Call results by funding instruments

Cooperative Research and Innovation Projects >Call 2019 >

Adaptive multi-modal network discovery for information assessment (AMMONIS)

The AMMONIS project develops services for the joint analysis of audio and text data using heterogeneous data sources. These services enable the classification and linking of audio events with text information that is spatially, temporally or thematically related. The aim is to use these services to find contradicting - or supplementary - news reports in the different news channels, and thus to support improved monitoring and a better assessment of crisis situations.

The acquisition, linking, assessment, and dissemination of information play an increasingly important role in modern conflict scenarios. In addition to classic military operations and economic or commercial pressure, so-called "hybrid warfare" additionally pursues political aims and disseminates (dis)information on the Internet and social media, for example to cause riots in the adversary's population or destabilize society.

The aim of AMMONIS is the development and implementation of prototypical services for the creation, adaptation and continuous evaluation of multi-modal AI algorithms for the classification and linking of audio events with text information that is spatially, temporally or thematically related.
The results of the project are modularized and reusable services for the linked analysis of audio and text data using heterogeneous data sources that can be provided in a closed cloud or HPC environment. Thanks to the combination of various modalities, the preparation of unstructured data, and the networking of information to form a semantic network, the services support the situation assessment.