The 2011 IS Challenge 

This year's challenge:  Exploiting Artificial Intelligence to Predict State Failures

In today’s world, there is a continuously available flow of reports describing newsworthy events at home and abroad. Traditional news media and emergent social networks offer a wealth of information.  Increasing inter-connectivity between states has raised Canada’s exposure to international turbulence.  In particular, fragile and failed states often require humanitarian assistance while at the same time risk becoming threats to Canada’s national security and stability. As a result, Canada has a vested interest in monitoring state tension and anticipating events that could trigger state failure. CAE, Carleton University, and Defence Research and Development Canada are investigating the feasibility of developing a predictive “early warning” model which would integrate structural (permissive) and events (precipitant) data.

Attempts have been made to characterize events reported in news media. This has proven problematic. Typically, efforts have relied on a single media source rather than the many sources of open data which are freely available. Perhaps more importantly, machine coding has failed to account for the context within which an event occurs. For example, an ethnic hate speech will have a very different impact in Canada than in a country with a history ethnic tension. Similarly, a series of car bombs in Ottawa would likely have greater perceived significance than a car bomb in Kabul. Currently, this contextual information must be interpreted by a human analyst. The challenge posed under this project is to exploit artificial intelligence and systems intelligence to automate this process and to represent the aggregated data visually.  

There are three key objectives to this challenge:

1. Develop an application that can monitor real-time, open source news feeds to categorize events.
2. Apply key performance indicators (KPIs) to characterize the event based on context.
3. Develop a software application that visualizes the data and draws attention to areas of rising state tension and warns of state failure.

A full description of the challenge problem and more information about the objectives is available below, along with a complete description of the data teams will be working with, and how programs submitted by contestants will be evaluated.

That challenge was proposed by CAE.  CAE is a world leader in providing modelling and simulation technologies and integrated training solutions for the civil aviation industry, defence forces, public security agencies, and healthcare organizations around the globe. With annual revenues exceeding C$1.6 billion, CAE employs more than 6,500 people at more than 75 sites and training locations in 20 countries. CAE’s Professional Services team provides analytical and engineering services, leveraging the use of modelling & simulation to support analysis, training, and operations.   

Prizes & Award Ceremony

The grand cash prizes for winning teams are intended to be totalizing 10,000 CAD, but the actual amount may be different depending on the IS Challenge budget. Awards will be officially announced at the 2011 Intelligent Systems Collaborative (AI/GI/CRV/IS) Conference in Saint John's, NL. Cash prizes will be given to the designated leader (or other designated representative) of each winning team. Winners for each category will be notified before the conference and are expected to attend the award ceremony at the conference. See challenge 2011 policies and guidelines for details.

How to Participate

1. Register your team.  If you are not ready to register, subscribe to the newsletter (right hand side of the screen) to be informed about any change, news or bug regarding the competition.
2. Download the problem description.  
3. Write a program or programs to address the proposed objectives.
4. Receive the test datasets.
5. Submit final entry, as described in the problem description file.