Gradiant takes part in this project providing a complete technology package to identify radicalization signs in social media
PRACTICIES is founded to mobilize networks of European cities with the purpose to better understand the “human roots of radicalisation” as a preventive measure
Mobilizing, connecting and creating a network of European cities and experts in different fields with the purpose to better understand the roots of violent radicalization is the objective of PRACTICIES initiative, a project funded by the European Union’s Horizon H2020, Gradiant participates in along together with other organizations.
After all the recent events happened connected to terrorism, EU aims with this kind of projects go to the origin of the real problem implementing actions that allow to identify these processes of “human roots of radicalisation” as a preventive measure.
It is essential to design specific tools for detection and this is exactly Gradiant’s role in this project. “Our mission is to develop and provide a package of complete technologies to identify signs of radicalization in social media, such as Twitter, Facebook or discussion forums and help other security experts determine what to do and how to act”, explains David Chaves, co-director of Intelligent Systems department at Gradiant.
PRACTICIES, named after the acronym ‘Partnership Against Violent Radicalisation in Cities’ brings together different agents to work towards a common purpose and contribute their knowledge and experience in their respective fields. This network includes national, regional and local institutions responsible for implementing policies to prevent radicalisation, as well as researchers, stakeholders, representatives of civil society and specialised technology organisations, such as Gradiant.
Natural language processing for an effective prevention
The solution developed by Gradiant uses text mining technology based on natural language processing (NLP). Chaves points out “we’re focused on giving structure to the human speech posted on the Internet to be understood by a computer, processed and aggressively identified in the published texts. With this information, which is tagged and classified, this tool offers a series of radicalization indicators to help PRACTICIES team, specialized in content supervision, to carry out their job. Tagging and content classification are the tasks we are focusing on, since it is the most tricky part of the software and the one essential to use it.
Moreover, Gradiant also provides computer vision technologies to automate image analysis to complete the tools for early identification or situations monitoring in order to give a useful and complete solution that really contributes to early prevention.
Expertise, key factor to provide fast and effective solutions
“It is important to have previous experience and developments in this and other technologies because it has made it easier for us to move forward quickly and safely in this European project”, highlights David Chaves. Gradiant have previously applied NLP technologies in content generation for marketing. In PRACTICIES project, the tool is adapted for online hate speech search in order to help the early detection and identification of people’s radicalization.
Intelligent Systems department at Gradiant is currently working on the internal organization of the system in order to provide a tool that can be used by other PRACTICIES members, who are security experts and do not necessarily have technical knowledge.
This first stage of system design and organisation is expected to be completed by the end of 2019 and then implemented in security policies to prevent violent radicalisation in European cities as a precautionary mechanism for terrorism. Within the framework of the European research project PRACTICIES, Kantar Public Brussels has conducted a ground breaking survey among 12,000 young Europeans on different issues which 47% of the young people interviewed consider that violent radicalisation is widespread in their country and 42% believe that this phenomenon will spread over the next few years. This highlights the relevance of this issue at an early age.