The Role of Computer Science in Civilian Casualty Recording and Estimation in Times of Conflict
Carnegie Mellon University
Organizers: Jay Aronson (Dept. of History) and Tom Mitchell (School of Computer Science, Machine Learning Department)
On Friday October 15, Carnegie Mellon University hosted a small workshop on the role of computer science in civilian casualty recording and estimation in times of conflict. In order to keep the meeting focused and productive, the organizers placed a particular emphasis on data acquisition, organization, and analysis. Five outside speakers were brought in to discuss the computing challenges they face in their work and to request specific guidance from faculty and graduate students from the School of Computer Science.
- Mike Spagat of Royal Holloway/University of London and a close collaborator of the Iraq Body Count spoke about the reasons why we should care about civilian casualties as well as methods of analyzing casualty data using agent based modeling.
- Jeff Klingner and Megan Price of Benetech’s Human Rights Data Analysis Group presented a series of computing and data-oriented challenges that they face in their work casualty recording and estimation in a variety of contexts. They also described a variety of other specific ways that the computer science community could contribute to the organization’s work, including the development of a cell-phone or smart-phone interface for their Martus human rights violation recording system.
- David Banks from Duke University discussed how demography and Agent-Based Modeling may enhance our understanding of civilian casualty data and conflict more generally. He presented several challenges to the audience in this regard in relation to Kurdish demography.
- Peter Sheridan Dodds from University of Vermont talked about the possibilities and pitfalls of using methods from computer science and theoretical mathematics to analyze complex social phenomena like violence and conflict.
By the end of the day, the general consensus among the CS faculty was that challenges described by practitioners could be categorized in three ways: problems with immediately implementable, off-the-shelf solutions; problems with existing solutions of varying quality that needed to be customized in order to meet the needs of the human rights community; and finally, problems that were at the frontier of computer science research or required original research to be solved. Because many of the challenges raised fit into the first two categories, computer science faculty and graduate students shared available techniques and resources with practitioners.