CAGE
Cultural Adversarial Game Engine
Cultural Adversarial Game Engine
The goal of the CAGE project is to develop a software platform within which we can rapidly recreate a given part of the world - complete with the "look and feel" of that part of the world and, more importantly, with people who behave in accordance with models of the behaviors of the socio-cultural-ethnic-religious groups to which they belong, and whose behaviors can be predicted through the use of some of our predictive tools such as SOMA and CONVEX (as well as others that may be developed by other researchers). Uses of CAGE can include training, as well as gaming "what if" scenarios with confidence that the underlying behavioral models and predictive algorithms have been rigorously tested and validated.
We are currently on the third iteration of CAGE-based games. Each application is aimed at a different user base, but shares much of the same underlying technology. At its heart, each game provides a method for its user to explore and interact with a hypothetical world containing one or more of the roughly eighty geopolitical actors the LCCD currently tracks.
Current CAGE Games
WebCAGE (2009): WebCAGE is aimed at the seasoned policy analyst interested in seeing how policy changes (e.g., stopping NGO support, legalizing a group, opening negotiations) will affect opposing actor responses (e.g., bombings, kidnappings) over an extending time period. The game itself focuses more on visualizing large datasets than creation of an immersive virtual world. As the name implies, this game is entirely web-based, though only available to select uses.
SAGE (2008): SAGE's target audience is the "newbie" policy analyst. We provide a 3D environment in which the user interacts with a significantly smaller subset of groups, actions, and environmental variables than in WebCAGE. By making changes to the hypothetical world in which these groups exist, the user discerns affects of policy changes and can use these to make better recommendations.
CAGE (2007): CAGE, our initial Cultural Island Game (CIG) was developed as a virtual world based on Second Life. In an Afghan village, a United States soldier must interact with residents ranging from the revered village elders to regular farmers. This game was built as a training simulation to reduce culture shock and prevent acting against societal norms.
Although the lab tends to track groups our sponsors find interesting (terrorist organizations), CAGE and the related predictive technologies found in engines like SOMA and CONVEX are theoretically proven to work on any data source. We are beginning to explore other, unrelated realms like disease tracking, and will post results as they manifest!
Project lead: Dr. V.S. Subrahmanian.
For additional information, please contact Dr. V.S. Subrahmanian or John Dickerson.
Last updated: November 2009 by John Dickerson.
Project Contributors
Special Thanks
Publications
This page will be updated as our work enters print. For information about receiving draft publications, technical reports, and conference presentations, please do not hesitate to contact team members.
The following sections may include links to restricted access material. Please do not hesitate to contact a group member for instructions regarding how to obtain a username and password.
Presentations
The following sections may include links to restricted access material. Please do not hesitate to contact a group member for instructions regarding how to obtain a username and password.
Live Demonstrations
The following links will take you to the various CAGE online portals.
Downloads
Links to related offline content:
Please get in touch with a group member if you would like access to the actual games!
WebCAGE
The following images give a brief overview of parts of the WebCAGE online environment.
After selecting the active groups Hamas, Harakat-e-Islami, Hezb-i Islami Khalis, and Hezbollah (offscreen), the user selects which action variables to include in the simulation.
Above, we see the user setting an environmental variable for Harakat-e-Islami. Specifically, the user is specifiying that Harakat-e-Islami does not receive foreign NGO support. This is setting the initial condition of our virtual world; variable assignment will affect how each group chooses to respond to player actions.
During one of the opposing groups' response phases, we see that three action variables have taken a turn for the worse (represented by red squares). One such variable, ARMATTACK, has increased by 1 since the last round. What policy recommendations made by our human user led to this increase?
Should the user choose to save and exit his game, we offer a post-game analysis option with automatic recognition of so-called "active variables" -- those variables that changed significantly over the course of the game. The interface pictured above offers a set of data parsing options, as well as easy exportation of parsed data.
SAGE
The following images give a brief overview of parts of the SAGE 3D gaming environment.
Some of the storyline portion of SAGE. Currently, the user is an analyst at a United States desk in Afghanistan.
Visualization of some of the user's recommendations regarding how to deal with a specific organization.
An Afghan insurgent weapons cache.
Consequences of a violent insurgent attack, stemming from poor user analyst recommendations.
An end-game recap of actions taken and their results over multiple time steps.
CAGE
Coming soon!