Collaborative Analysis & Knowledge Exploration (CAKE)

The CAKE solution is a semantic framework that supports information discovery, sense making, and presentation in a dynamic, collaborative environment. The technology reduces ‘data-to-decision’ time through the use of semantic and collaborative visual analytics techniques.

Today’s defense and security analysts are drowning in a flood of data: modern sensors and reporting systems are producing unprecedented amounts of data, and modern processing capabilities are transforming and translating this data at an ever-increasing rate. Tomorrow’s defense and security analysts must be able to find, corroborate, and interpret actionable information in this data flood in an efficient and timely manner.

Key to achieving this goal is the capability for collaboration. Emerging information technologies and improvements in communications capabilities make collaboration in near real time a reality today. New methods for analyzing and assessing data, including ontology modeling and data visualization techniques, mean that analysts have more to collaborate with: the ability to better analyze and assess data provides analysts with more observations, hypotheses, and findings to communicate.

In Phase I of the Collaborative Analysis and Knowledge Exploration (CAKE) initiative, KBSI investigated the design of a semantic framework that supports information discovery, sense making, and presentation in a dynamic, collaborative environment. The Phase I work established the requirements for collaborative sense making, formulating techniques for fusing tagged text and image data, and designed a method for collaborative sense making with enhanced situational awareness.

The CAKE innovations include (i) an intelligent, ontology-driven approach for automated semantic tagging of text data; (ii) a new ontology-driven method for effectively fusing information from text, image, and geospatial data; (iii) advanced semantic search and discovery from large multi-source tagged data collections; and (iv) new methods for interactive visualization to support collaborative sense making, leading to enhanced situational awareness.

Phase II Development

The Phase II CAKE initiative is establishing and validating the technology for semantic knowledge creation from multi-modal text and image data feeds via semantic tagging and collaborative visual approaches. The initiative is currently focused on refining the CAKE method, investigating data sources, hardening the needs and requirements of the CAKE application, and leveraging and adapting components of KBSI’s Threat Assessment Dashboard (THAD™), particularly the THAD™ technology’s Bayesian belief network, which is used to calculate a threat index by fusing a variety of factors that might impact the potential of an emerging threat.

The CAKE method includes four functions: semantic tagging, information fusion, discovering knowledge for sense making, and providing collaborative visual analytics. In semantic tagging, CAKE labels text and image data with tags that provide meaning in the context of applications related to the information in the data. KBSI is using an ontology to increase the semantic quality (depth of meaning represented) of the text and image tags.

The semantic tagging activity produces an integrated tagged data set that combines the semantic information contained in the tagged text, tagged images, and geospatial databases. The dataset provides an expressively rich knowledge base that is useful for exploratory searches and collaborative visualization and sense making. Using the tagged data, the CAKE information fusion approach uses an application domain ontology to intelligently guide the focused contextualization of the fusion activity. The ontology acts as a bridge that helps determine the mapping between information items in the different tagged data forms (text, images, and geospatial data).

CAKE uses several mechanisms to collaboratively ‘discover’ action-enabling knowledge from this information: (i) semantic search, (ii) social network extraction and analysis, and (iii) event extraction and analysis. CAKE’s dynamic visualization mechanisms enable collaborating end users to gain a better understanding of information contained within the data. CAKE uses a semantic tag-based approach that maintains and traces through hierarchies of different scales of space and time granularity for asynchronous collaborative visual analytics. The semantic tags, carrying scaled spatio-temporal information, helps users navigate between text visualization, space visualization, temporal visualization, and spatio-temporal visualization.

The CAKE technology will allow intelligence analysts to answer questions such as what or who are the objects and persons of interest in the emerging situation, who are the key leaders and what are their social networks, where are the events taking place, how are the events being carried out, why are the operations of interest, etc. CAKE allows analysts to answer these questions by augmenting human subject matter expertise with fused intelligence derived from multiple sources and sensors while taking into consideration established, expert knowledge.

The CAKE solution will provide a number of benefits, chief among these is a reduction in ‘data-to-decision’ time through the use of semantic and collaborative visual analytics techniques.  By fusing multiple data sources, the technology will also provide substantial gains in the quality of shared situational awareness and significant increases in the ability to exploit information and knowledge embedded in the data.

Related Research

Related Solutions