Officials of the Intelligence Advanced Research Projects Agency (IARPA) in Washington have issued a request for information (IARPA-RFI-14-05) on an upcoming program called Emerging Events and Participating Entities: Knowledge Representation and Reasoning with Uncertainty and over Time.
IARPA officials are trying to determine if such a program is feasible to develop computers and software that can identify and track unfolding events before a big picture becomes clear. IARPA is the research arm of the Director of National Intelligence.
IARPA experts eventually may issue a formal program solicitation if they receive promising information from industry.
Maintaining national security involves identifying emerging and evolving events, IARPA researchers explain. Given the sea of information that analysts must navigate daily, identifying activities of interest is a huge challenge.
Related: IARPA seeks to move high-performance computing beyond CMOS and floating point
Current technologies do a reasonable job of identifying facts, but have trouble identifying the events that link these facts together, researchers say. Some technologies have tried to identify events in data by identifying related verbs, but have met with little success.
The ability to identify and track events over time could help intelligence analysts understand unfolding events and the people and equipment involved, IARPA officials say. Doing this is difficult, however, because unfolding events often are unclear.
Big events like a presidential inauguration are straightforward because they have clear elements. Precursor events take place long before the major event. Planning takes place; permission is obtained; decisions are made; purchases are made; participants come and go. These events also link to other events.
Many events that turn out to be significant, however, are difficult to identify -- especially when their goals are not apparent. Identifying and understanding time and reasoning about the emergence of events and the sequencing of activities can be extremely difficult, IARPA officials say.
From industry, IARPA wants opinions on whether it's possible today to make significant progress on tracking this kind of event emergence. Intelligence analysts need ways to define significant emerging events clearly, and track events and the people involved as events unfold. Analysts, furthermore, need to move beyond verb-based event definition to approaches that find and link events as big pictures start to emerge.
Automating how analysts identify and track emerging events and the people involved must be able to define significant events, precursor events, and emerging events; identify people, places, and equipment involved in emerging events; draw conclusions about events over time; and reason about uncertain emerging events.
From industry, IARPA researchers are interested in a discussion of technologies that support identifying emerging events, their components, and the people involved and that can reason about the events over time.
Related: IARPA seeks industry help with artificial intelligence approaches to interpreting data
IARPA researchers will use information from industry to help them plan a one-day workshop this June in the Washington area that focuses on new approaches to understanding, characterizing, and capturing emerging events.
Companies interested should email responses formatted as .pdf documents no later than 28 March 2014 to IARPA's Heather McCallum-Bayliss at [email protected]. Email questions or comments to McCallum-Bayliss at [email protected].
More information is online at https://www.fbo.gov/notices/94d19f6d51395999e47a69fd4bbfb977.
