Fast Forward: Forecasting Global Emerging Threats
Resources and FAQ for the IDEaS contest “Fast forward: Forecasting global emerging threats”, hosted by the Canadian Department of National Defence and the Canadian Armed Forces (DND/CAF).
This challenge is hosted and administered by the Canadian DND/CAF. Please consult the official contest page and personal instructions from the IDEaS contest team for official data and application material. This page only
serves to provide a preliminary outline of the training data, and answers to frequently asked questions. We kindly ask that any further questions are directed to the contest team at
IDEaSContests.IDEeSConcours@forces.gc.ca.
About the challenge
The task
The Canadian Department of National Defence and the Canadian Armed Forces (DND/CAF) invite research teams and innovators across the globe to take on the new
Fast Forward – Forecasting Global Emerging Threats prediction challenge. Offered through the Innovation for Defence Excellence and Security (IDEaS) program, it calls
for expert teams to harness the power of publicly available data, statistical methods, and machine learning, to build forecasting models that can provide forecasts for fatalities in future armed conflicts,
accompanied by machine-generated explanations of how the models work.
Up to ten qualifying teams will be invited to proceed in the contest following the application deadline on 10 December 2024. They will receive a $25,000 development grant to
create their models.
Awards for the top three teams
Based on the performance of their respective predictions, and the intuitive documentation of their final models, the top three teams will be awarded prizes of up to $150 000. Their contributions will also help advance the use of data-driven
forecasting methods by Canada’s Defense Team:
“The longer-term goal is to integrate these models into methods and tools aimed at helping the Canadian Defence Team intelligence analysts to perform data-driven forecasting tasks while addressing the challenges of trust and accountability for analysts leveraging these novel forms of anticipatory intelligence. The aim will be to compare and eventually adopt systems that can help analysts to rapidly organize and exploit data from publicly available and private sources.”
(canada.ca)
Eligibility & Application procedure
Who can apply?
Please note that applicants must be legal entities duly incorporated and validly existing in Canada, including:
- Canadian universities and educational institutions chartered in Canada.
- Incorporated Canadian not-for-profit organizations or associations.
- Incorporated Canadian for-profit companies, organizations, or associations.
- Provincial/territorial, or municipal government organizations. Federal and provincial crown corporations are not eligible for funding.
How to apply
The deadline to apply is 10 December 2024.
For instructions on how to apply, please email the IDEaS Contests Team at
IDEaSContests.IDEeSConcours@forces.gc.ca.
TRAINING DATA
What does the training dataset capture?
A subset of the comprehensive input dataset that informs the Violence and Impacts Early-Warning System (VIEWS) will serve as the training and calibration dataset for the applicants’ models. It is comprised by spatially and temporally granular data from trusted open data providers, such as UCDP, V-DEM, and the World Bank (WDI). Adapted to the country-month and PRIO-GRID-month levels of analysis, it captures themes ranging from conflict history to development and democracy indices, social and natural geography, climate indices, conflict-related topics in news media, and much more.
List of features at country and grid level
A complete list of the features that will be provided at the country and grid levels an be found below (as pdf).
Sample data
An example of what the data will look like is available via the link below, showing the training dataset provided for the 2023/2024 VIEWS Prediction Challenge. The dataset to be used for the IDEaS challenge will contain an updated version of this dataset, complete will data from 1990 up to the month prior to the first month that applicants are asked to forecast.
GIS resources
What are the units of analysis?
The forecasts should be provided at the country-month and PRIO-GRID-month levels of analysis, matching the units used by the VIEWS mode.. Information about each of these units, including references and shapefiles, are provided via the links below.