Submission of Internal Report to UN Partners
An interdisciplinary research group at ETH Zurich submitted an internal report to the UN Operations and Crisis Centre.
In September 2023, an interdisciplinary research group composed of computer scientists and global peace and security scholars submitted an internal report to the UN Operations and Crisis Centre (UNOCC). The report explored how conflict-event data from the UN’s ‘Situational Awareness Geospatial Enterprise’ (SAGE) database could be combined with machine learning. The report was the outcome of an 18-month pilot project led by the ETH Center for Security Studies (CSS) and realised in collaboration with the ‘Information Science and Engineering’ group at the ETH Institute for Machine Learning. The report ‘Towards Conflict-Event Prediction for UN Peacekeeping: Pilot Results based on SAGE and Avenues for Future Work’ summarised the pilot project’s main findings for UN policymakers and practitioners.
Key Takeaways:
- The report proposed a method by which deep neural networks can be combined with SAGE to generate short-term forecasts of conflict intensity across regions with a UN peacekeeping presence.
- Applying the proposed method to data from three large UN peace operations yielded encouraging results, but also showed the need for a more extensive mapping of the potential design space.
- The report argued that research on predictive methods for UN peacekeeping should eventually pursue a ‘collaborative’ approach in which different machine-learning models and human experts can make conflict forecasts together.
