STAR: Semantic Temporal Association Retrieval

Intelligence analysts gather and synthesize information from multiple sources in response to both high priority requests and standing topics. Their expertise enables them to evaluate and process information effectively. However, the time they spend searching for relevant information inevitably subtracts from the time available to develop analyses, especially predictions, from those data. As ops tempo accelerates, information overload will only increase; the time spent searching will grow while the time spent analyzing intelligence will decline. The STAR (Semantic Temporal Association Retrieval) system is designed to reverse this trend. STAR will support information triage, analysis, and discovery. It will enable analysts to allocate more of their time to analysis, and directly support them individually and collectively in defining trends and making predictions. STAR combines (1) the latest techniques in latent variable methods for digesting masses of textual data with (2) temporal analysis techniques for making predictions. It uses these methods to automatically compare recent and historical patterns, and support analysts as they judge the probability of future events. STAR will also provide components of a software infrastructure that will support richer and more effective collaboration among analysts by enabling them to query and build upon each other’s results.