Now under the sponsorship of the Office of Naval Research (ONR), the goal of W-ICEWS is to extend the ICEWS system to cover countries in all the major geographic combatant commands. SAE is concentrating on refining and extending its instability forecasting models to improve accuracy, transparency, and operational utility.
SAE is a prime contractor on SIMPL, which is sponsored by the Office of Naval Research (ONR). The goal of the program is to develop a capability to automatically parse and geo-locate dynamic events data at the sub-national level, and use those data (along with other sources of data) to generate reliable forecasts of local-level conditions and dynamics that could give rise to violence and instability. Local-level forecasts of instability are useful to military planners and operators who need to prioritize the application of stability resources and activities to specific areas where they are most needed and most likely to make a difference.
SAE, as a prime contractor to ONR, is developing a next-generation text processing capability to automatically extract multi-adic emotions, sentiments, and opinions from electronic sources of text. Since emotions can strongly influence behavior, the goal is to help decision makers better understand how their actions can mitigate the intensification of violent political conflict and simultaneously aid reconstruction and development operations.
SAE was a major sub-contractor on the Defense Advanced Research Projects Agency's (DARPA) ICEWS program. The goal of ICEWS was to develop and integrate multiple data feeds and analytic models to generate reliable forecasts of violence and upheaval around the world. Since its inception in 2007, ICEWS has been focused on the PACOM AOR (with plans/interest to expand throughout the world) and ICEWS is engaged in software integration exercises with STRATCOM's ISPAN program. SAE contributed to three key components, which transitioned for operational use at USPACOM, USSOUTHCOM, and USSTRATCOM:
- iTrace: uses a suite of natural language processing (NLP) algorithms to quickly and automatically transform unstructured news reports into structured quantitative indices that measure who is doing what, to whom, both between and within countries of interest. The customizable indices can be layered and visualized in a web-based interface that provides a 360 degree view of huge volumes (>10 million) of news reports.
- iCast: uses data from iTrace, along with other sources of data, in quantitative models to continuously generate forecasts of a variety of events of interest, including domestic political crises, ethnic/religious violence, rebellions, insurgencies, and international crises. These forecasts have so far demonstrated about 80%+ accuracy.
- iSent: automatically parses blogs and newsfeeds to extract quantitative data that reflect who is saying (or feeling) what about whom, or what other organization, issue or activity that was conducted in a particular region (i.e., provision of humanitarian aid). iSent is an efficient alternative to traditional public opinion polling.
SAE consulted on various Strategic Multi-Layer Assessment (SMA) program projects which is part of the Joint Staff/J-3, STRATCOM/GISC, and the Rapid Technology Program Office within the Department of Defense Research and Engineering. In particular, SAE analysis was used in reports on assessing the ability of the US to anticipate rare events which specifically referred to catastrophic terrorist events, including the use of a weapon of mass destruction. SAE analysis focused on an assessment of global and regional nuclear smuggling activities.
SAE is involved with a larger set of researchers on the Pathways project via the Combating Terrorism Technology Support Office's HSCB program. The goal is to develop an integrative system that monitors, models, and produces analysis for operators interested in specific events of interest. In particular, SAE will provide counterfactual modeling capabilities as part of this program to better understand how specific US actions affect political stability.
CAPES specifically focuses on assessing shifts in behavior of violent or potentially violent groups for the Air Force Research Laboratory. SAE collected and analyzed data on particular groups activities. SAE then used such data in statistical models to explain and forecast shifts in violent and nonviolent behavior in various groups.