N141-071 TITLE: Information Triage Engine: User-Composable Automation for the Dynamic Management of Text-Based Information

TECHNOLOGY AREAS: Information Systems, Human Systems

OBJECTIVE: Develop a scientifically-principled design specification with supporting prototype concepts for a set of advanced, user composable information management tools. Tools would enable warfighters to easily examine message traffic, email or other data streams of text based information. It would also enable warfighters to categorize, meta-tag, alert or otherwise manage information transactions based on natural language-like rules to support the execution of operations in the face of evolving missions with dynamic tasking requirements. Result will be a structured approach to user-composable automation that enables warfighters to manage information dynamically given real-time, emergent, missions and tasks.

DESCRIPTION: Many military domains require agile, time-critical decision making in order to maintain speed of command. These domains increasingly involve diverse and/or complex missions with dynamic information needs. Current information systems often get in the way of decision-making because they are data-centric, vice task-centric, and as a result require human decision-makers to manage large volumes of data before they can address command tasks. Warfighters are required to sort out what data is relevant to emergent tasks and mission demands. Tools are needed that allow decision-makers to rapidly configure (or compose) automation for managing the flow of data from multiple sources as needed to support emergent missions and tasks. The tools should support publish and subscribe (pub/sub), services oriented architecture. This effort will result in the natural language management of modular, reusable, and robust automation to be used with a variety of databases and information processing tools.

Given the demands of modern missions, automation must be easily composed by warfighters through natural behavior and language mechanisms that do not require advance programming skills. Warfighters need to be able to demonstrate the desired information management behaviors to the automation, and express them through natural language transactions and/or a graphical user interface. The desired Operator Machine Interfaces (OMIs) need to be task-centric and support the task needs of decision makers. The desired solution will result in an extensible strategy with supporting information services that: 1) facilitate rapid, behaviorally based, composition of information management in response to missions and tasks; 2) be applicable to a wide range of information feeds and data; 3) require minimal user training, 4) allow rapid insertion into DoD and Navy Information Technology, C2 and combat systems, and 5) demonstrate rapid end user adoption and utilization of the tools developed in accordance with the design schema. Resulting products will support improved operational (e.g. combat system; command and control (C2)) decision making and reduced workload by providing a dynamic, interaction-based mechanism for managing data based on mission and task information context. The desired tools are expected to enable command staff to more rapidly plan, re-plan and execute missions in the face of changing mission requirements and information needs. The desired solution would represent an approach to a reusable Information Triage Engine.

Desirable features would include:
� Natural interface (e.g. one which is behaviorally trained) for composing a workflow engine in response to task and mission demands.
� Rapidly composed through natural language and/or behaviors.
� Include mechanisms to support the dynamic tagging of information with appropriate meta-tags
� Incorporate mechanisms to support iterative tuning of information management engine processing rules.
� Consider mechanisms to refine automation to include negative (not) information, (e.g. to exclude certain categories of information) as well as other relational arguments.
� Support for automated derivation of meta-tags based on mission & task context and user activity.
� Demonstrate the generation of alerting rules for (proactively) alerting the user to relevant information and/or changing needs based on new/upcoming events or text (i.e. context shifts)
� Consider mechanized standardized information summaries. These might include template-based summaries for what information is/i not available in (and/or is being aggregated into) data stores; as well as context descriptors of time, location, and related mission(s) for which the information was collected.
� Demonstrate an ability to store user-composed automation as rule sets for discovery & reuse.

PHASE I: Phase I will address an approach for managing a stream of text-based messages from at least one application/database. Develop one or more use cases for how your proposed system will assist a decision maker in the dynamic management of information. Identify relevant literature from cognitive decision making and human supervisory control of automation to develop requirements for decision support needed for the proposed technology. Design a concept prototype tool with wireframe design elements to demonstrate interactions between human supervisor and anomaly detection technology. Define operational and technical metrics that will permit the demonstration of the utility of the approach in Phase II. Propose notional elements for an extensible decision support mechanisms by which warfighters might compose automation for these elements that could then be dynamically adapted to support mission execution and re-planning tasks. Design and prototype a basic proof-of-concept decision support prototype capability that would demonstrate improved speed of command required to perform re-planning. Demonstrate through storyboards or a working prototype how users would create, select and modify composable automation as needed, including select from available algorithms and technologies. Define and analyze the requirements for determining mission and task management requirements, and mission context for a specific information triage task. Show how automation would be managed for reuse by future users and with other applications/databases. Phase I deliverables should include a Final Phase I report that includes a detailed description of the approach taken, as well as a detailed proposed approach for Phase II development.

PHASE II: Develop, demonstrate, and refine the Phase I concept prototype. Validate utility in human performance based evaluations. Demonstrate applicability to an operational domain appropriate for the planned commercialization strategy (e.g. command and control). The effectiveness of the system shall be demonstrated by satisfying the utility metrics defined in Phase I, as well as additional metrics that may be developed in Phase II. Develop a formal plan for transition and commercialization.

1. Mature, demonstrate, and refine a concept prototype illustrating the decision support concept.
2. Phase II demonstration to address information orchestration across multiple applications/databases. Databases may include e-mail, naval message traffic, social media, etc.
3. Validate the decision support concept and quantify its impact on decision making tasks with controlled human performance study with dynamically changing mission/task requirements with empirical, user-based performance data appropriate to the anticipated transition end-user.
4. Demonstrate that the schema can accommodate a variety of different decision making tasks.
5. Develop a transition strategy for insertion of the technology developed into a US Navy Program of Record (e.g. BYG-1, GCCS-M) or a commercially developed system.

Phase II deliverables will include a Final Phase II report that includes a detailed description of the approach taken and results obtained in Tasks 1-5, and formal transition agreements or commitments for external investments as part of a Phase III development effort.

PHASE III: Refine the prototype and make its principal features complete in preparation for transition and commercialization. In addition to the DoD, there will be an increasing demand for supervision of autonomous systems in the commercial sector, such as the process control domain and commercial mining industries, and in federal and state agencies such as law enforcement, emergency management, and border protection. These domains could benefit significantly from the application of the solution developed in this effort. In operational command centers, it is expected that the developed tools would serve to create user composable automation for critical mission events that would be applicable to Fleet command centers and shipboard combat centers supporting anti-submarine warfare, or Area Access Denial missions.

PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: There is an ever increasing demand for user-composable, readily extensible, automation that enables the dynamic management of information. Tools for detecting, editing and responding to the concurrence of data from disparate sources is endemic to modern life. The desired technology will have ready application to streaming web data accessed through computers and personal electronic devices. The desired technology would have ready application to next generation office automation providing productivity gains to individuals and across organizations. Composable automation algorithms would create a new form of autonomous systems for information management. Further, these tools would become key enablers to the supervision of autonomous systems in the commercial sector, such as in the process control domain and commercial mining industries, and in federal and state agencies such as law enforcement, emergency management, and border protection. Multiple application domains could benefit significantly from the application of the solution developed in this effort.

REFERENCES:
1. Hollnagel E (1993) Human reliability analysis: context and control. Academic Press, London.

2. Parasuraman, R., Sheridan, T. B., & Wickens, C. D. (2000). A model of types and levels of human interaction with automation. IEEE Transactions on Systems, Man, and Cybernetics Part A: Systems and Humans, 30, 286�297.

3. St. John, M., Smallman, H.S., Manes, D.I., Feher, B.A., and Morrison, J.G. (2005) Heuristic automation for decluttering tactical displays. Human Factors, 47, 509-525.

KEYWORDS: Dynamic decision making, natural language, command and control, context modeling, information management

** TOPIC AUTHOR (TPOC) **

DoD Notice:   Between November 20 and December 19 you may talk directly with the Topic Authors (TPOC) to ask technical questions about the topics. Their contact information is listed above. For reasons of competitive fairness, direct communication between proposers and topic authors is not allowed starting Dec 20, 2013, when DoD begins accepting proposals for this solicitation.

However, proposers may still submit written questions about solicitation topics through the DoD's SBIR/STTR Interactive Topic Information System (SITIS), in which the questioner and respondent remain anonymous and all questions and answers are posted electronically for general viewing until the solicitation closes. All proposers are advised to monitor SITIS (14.1 Q&A) during the solicitation period for questions and answers, and other significant information, relevant to the SBIR 14.1 topic under which they are proposing.

If you have general questions about DoD SBIR program, please contact the DoD SBIR Help Desk at (866) 724-7457 or email weblink.