CUPID: Cognitive-service Utilized for Proactive Intelligent Decision-support Navy STTR FY2013.A Sol No.: Navy STTR FY2013.A Topic No.: N13A-T020 Topic Title: CUPID: Cognitive-service Utilized for Proactive Intelligent Decision-support Proposal No.: N13A-020-0164 Firm: Aptima, Inc. 12 Gill Street Suite 1400 Woburn, Massachusetts 01801 Contact: Nathan Schurr Phone: (781) 935-3966 Web Site: aptima.com Abstract: The Navy's mission success increasingly relies on complex decision support systems and automation. Navy command and control systems must provide the users with the right information, at the right time, depending on the context in which they are operating, and must do it in a proactive manner. Today, massive quantities of data are passed blindly between nodes, with no ability to adapt the delivery or presentation of information to suit users' needs. To address these challenges, Aptima, with the support of Colorado State University (C.A.P. Smith) and consultant Krishna Pattipati, proposes to develop a Cognitive-service Utilized for Proactive Intelligent Decision-support (CUPID). The Aptima team will ensure success by leveraging a three pronged effort including: 1) groundbreaking theoretical design approach and metrics; 2) rapid prototyping of a tangible Cognitive Service middleware; and 3) pursuing a clearly identified transition environment, the Distributed Common Ground System-Navy (DCGS-N). The Aptima team believes that the CUPID Cognitive Service will offer unique benefits to C2 system designers and users by enabling productive, proactive user interactions. Benefits: CUPID will create a rich cognitive service module that enables more productive, proactive interactions with the user by combining an existing context model to capture cognitive state with reasoning about optimal proactive system interactions. To leverage the context information and enable more accurate and understandable interactions with the user, a reasoning component will initially focus on two challenges in decision support systems: uncertainty and interaction method. Colorado State Associate Professor C.A.P. Smith will be providing insights on decision support design specifications, optimal interaction methods, and validation (see Adaptive Supervisory Control in Related Work below). Consultant Krishna Pattipati will be supporting the quantification and propagation of uncertainty through the decision support system. In addition, we anticipate that the need for training on a given decision support system will likely be reduced, because CUPID supports the current cognitive state and context of the user and adapts the proactive action based on that context. CUPID will provide a rapid, affordable prototyping capability by leveraging existing tools and frameworks such as CCRF: Common Context Representation Framework (Ganberg et al., 2011). CCRF represents context on five major dimensions: (1) the user (whether human or machine); (2) the mission (including tasks, goals, and constraints); (3) the environment (including both the physical world and the enemy); (4) the interactions required for the mission (including both information and resources); and (5) the domain concepts necessary for a given application. In addition, it will be important for CUPID to adhere to standard APIs for integration. CUPID will be developed in a reusable Service Oriented Architecture (SOA) to enable its efficient application to future systems, allowing for rapid insertion of decision support concepts into current and future C2 & combat systems. CUPID has a clear transition path to DCGS-N Increment 2, in which it will serve as a cognitive service as part of the SOA architecture in the CANES system. The Aptima team has already met several times with BAE, the prime on the current DCGS-N Increment 1 effort, and they have offered to help shape CUPID and present it to the PMO for potential Increment 2 integration. The Aptima team plans to maintain communications with BAE throughout the Phase I effort, to increase the potential for end user adoption and utilization of CUPID. Return