Defense Advanced Research Projects AgencyTagged Content List

Analytics for Data at Massive Scales

Extracting information from large data sets

Showing 73 results for Analytics RSS
Dramatic success in machine learning has led to a torrent of Artificial Intelligence (AI) applications. Continued advances promise to produce autonomous systems that will perceive, learn, decide, and act on their own. However, the effectiveness of these systems is limited by the machine’s current inability to explain their decisions and actions to human users (Figure 1). The Department of Defense (DoD) is facing challenges that demand more intelligent, autonomous, and symbiotic systems. Explainable AI—especially explainable machine learning—will be essential if future warfighters are to understand, appropriately trust, and effectively manage an emerging generation of artificially intelligent machine partners.
The threat of distributed denial of service (DDoS) attacks has been well-recognized in the data networking world for two decades. Such attacks are orchestrated by sets of networked hosts that collectively act to disrupt or deny access to information, communications or computing capabilities, generally by exhausting critical resources such as bandwidth, processor capacity or memory of targeted resources.
The Geospatial Cloud Analytics (GCA) program is developing technology to rapidly access the most up-to-date commercial and open-source satellite imagery, as well as automated machine learning tools to analyze this data. Current approaches to geospatial analysis are ad hoc and time intensive, as they require gathering and curating data from a large number of available sources, downloading the data to specific locations, and running it through separate suites of analytics tools.
The social sciences can play important roles in assisting military planners and decision-makers who are trying to understand complex human social behaviors and systems, potentially facilitating a wide range of missions including humanitarian, stability, and counter-insurgency operations. Current social science approaches to studying behavior rely on a variety of modeling methods—both qualitative and quantitative—which seek to make inferences about the causes of social phenomena on the basis of observations in the real-world. Yet little is known about how accurate these methods and models really are, let alone whether the connections they observe and predict are truly matters of cause and effect or mere correlations.
Social media, sensor feeds, and scientific studies generate large amounts of valuable data. However, understanding the relationships among this data can be challenging. Graph analytics has emerged as an approach by which analysts can efficiently examine the structure of the large networks produced from these data sources and draw conclusions from the observed patterns.