VOX-Pol Pre-Conference Workshop 3 – Human Reasoning Meets Machine Learning


Human Reasoning Meets Machine Learning: Hands-on Workshop Using Mixed-Media Analytics Tools on Violent Online Extremism Data.

Date and time

Wednesday, 22 June from 2:00pm to 6:00pm


Heather Griffioen-Young – A social psychologist and manager of TNO’s Dark Web programme, Dr. Griffioen-Young has been involved in multiple projects involving radicalisation and violent extremism. Recent focus is on what we can learn about radicalisation and terrorism from online activity.

Technical Contributors

Serena Oggero holds a PhD in particle physics from the VU University of Amsterdam (NL) for data-analysis work at the CERN laboratory. Serena’s research was part of the LHCb effort to search for a rare “beauty” decay. Since 2013 Serena works at TNO (Netherlands Organisation for Applied Scientific Research), currently on data analytics (AI) innovations for Intelligence & Security. Serena’s driving question is: how to make data innovations “actionable” for their future users?

Stevan Rudinac is a postdoctoral researcher at the Informatics Institute of the University of Amsterdam. In his research he aims at enabling multimedia information retrieval based on the relevance criteria defined at a high semantic level, by jointly analysing visual content and the heterogeneous information associated with it, ranging from text and automatically generated metadata to information about users and their social network.

Workshop Descriptor

Much has been said about the online availability of data concerning violent political extremism. We believe mixed-media analytics tools can empower the collection of relevant content and effectively support those researching violent online political extremism. There’s also a need however for a common experimentation ground to explore where exactly mixed-media analytics provide a new grip on researchers’ questions, and how these questions translate to a man-machine environment.

– Would you like to discover which violent online extremism data can be disclosed from open and hidden online networks?
– Would you like to interrogate these datasets, to play with different visual and textual sources and to explore which issues in violent online extremism can be tackled from a data-driven perspective?
– Would you like to explore how your usual research approach can be complemented with state-of-the-art machine learning tools?

If so, this workshop is for you.

During the half-day session, we will:

– Introduce you to our multi-modal analysis approach on a broad range of different data (e.g. text, still images, videos) collected from various online information sources (e.g. online forums, social media platforms, webpages);
– Give you the opportunity to “interrogate” some of the extracted data sets and to test your questions with our software tools;
– Discuss our interfaces and visualisations, and their impact on your analysis work. We invite you to share your ideas with us to shape the tools and how best to adapt them to your work approach.

Who is discussing what types of concepts? What typical images are linked to specific concepts and why? Can we find recurring trends in communication styles or unusual online behaviors? These are some of the questions we expect to address during the interactive session and that we hope you want to test with us.

What You Need to Bring

A laptop + curiosity, will to experiment, and your relevant research questions!



2:00 pm - 6:00 pm
Map Unavailable