AI-Based Trend and Sentiment Analytics for Latent-Risk Discovery
Navy SBIR 2019.2 - Topic N192-131
ONR - Ms. Lore-Anne Ponirakis - email@example.com
Opens: May 31, 2019 - Closes: July 1, 2019 (8:00 PM ET)
TECHNOLOGY AREA(S): Human Systems, Information Systems, Sensors ACQUISITION PROGRAM: Minerva INP
The technology within this topic is restricted under the International Traffic in Arms Regulation (ITAR), 22 CFR Parts 120-130, which controls the export and import of defense-related material and services, including export of sensitive technical data, or the Export Administration Regulation (EAR), 15 CFR Parts 730-774, which controls dual use items. Offerors must disclose any proposed use of foreign nationals (FNs), their country(ies) of origin, the type of visa or work permit possessed, and the statement of work (SOW) tasks intended for accomplishment by the FN(s) in accordance with section 3.5 of the Announcement. Offerors are advised foreign nationals proposed to perform on
this topic may be restricted due to the technical data under US Export Control Laws.
OBJECTIVE: In support of forward deployed operations to protect soldiers, sailors, airmen, and marines operating in a coalition environment against insider threats, develop multidimensional machine learning and reasoning technologies that incorporate trend and sentiment analysis techniques and algorithms into a range of entity and behavior analytics for integration into a shared-networked environment for timely intervention and neutralization of harmful intents. This Artificial intelligence (AI)-based Decision-Aid technology is aimed at isolating and marking susceptible entities/groups that are willingly influenced by like-minded role-models, and may act on perceived cues to harm or terrorize. The marked entities/groups-of-interest are guided by ideological attitudes and perceptions encapsulating their set of values and interpretation of the world. To “connect the dots”, this AI technology exploits resources such as: open-source intelligence, social and financial network activities, and entities' stability to discover, identify, and predict the evolving dark-pattern-of-life that is accentuated by emergent behaviors associated with risk- latent intents—especially the risk indicators and warning signs related to low-signal-to-noise events and transactions are of particular interest.
DESCRIPTION: AI-based trend analysis looks for patterns or trends in the way that information changes and can be used to anticipate events or behaviors. Sentiment analysis is the process of analyzing multitudes of evidential transactions and salient-signatures (from voice to text to financial to social network) to assess entities’ attitude and emotional states. Change in sentiment measures for an entity over time can reveal evolving behavior and more importantly of the emotional state and the intensity. Performing automatic trend analysis on evolving behaviors can be used as a tracking mechanism to trigger alerts. This process can be used to understand and profile entities of interest or groups of entities and continually model their evolving behaviors and predict intent. Current techniques and tools are hand crafted using subject matter experts, often based on ad-hoc insights, and do not scale. The accuracy of information and resulting interpretations requires drawn-out independent assessment and are not practical for real-time operations.
This SBIR topic seeks the design, development, and demonstration of a prototype for open scalable architecture and AI-based multidimensional-trend analytics and learning methods that can exploit behavior analysis techniques and provide insight into the entities’ changing pattern of life. The proposer will develop AI methods to understand and profile susceptible entities or groups of interest by continually modeling their evolving behaviors and predict their intent in context as to affecting entities’ stability and the state of perception that things are changing, or have changed, or will change over time. The proposer will develop automated detection techniques for identification and tracking of the low-signal-to-noise indicators, which can be used as tags for monitoring and alerting aberrant activities and behavioral dynamics in the native environment; and also to detect and monitor changes in those activities or flag emerging activities. In other words, the proposer will develop learning algorithms for complex behaviors, their aggregates, and reciprocal behaviors when a subject engages in certain but limited social network and business transactions. The proposer will develop a prototype that performs: a) object discovery and tracking, b) intent discovery and tracking, c) social network interaction discovery and tracking, and d) procedural/business transaction process discovery and tracking. The prototype will incorporate or supplant existing state-of-the-art techniques being implemented by both the Intelligence Community and commercial sector. Proposed solutions can take advantage of existing social media data sources and emerging cultural behaviors.
Work produced in Phase II may become classified. Note: The prospective contractor(s) must be U.S. owned and operated with no foreign influence as defined by DoD 5220.22-M, National Industrial Security Program Operating Manual, unless acceptable mitigating procedures can and have been implemented and approved by the Defense Security Service (DSS). The selected contractor and/or subcontractor must be able to acquire and maintain a secret level facility and Personnel Security Clearances, in order to perform on advanced phases of this project as set forth by DSS and ONR in order to gain access to classified information pertaining to the national defense of the United States and its allies; this will be an inherent requirement. The selected company will be required to safeguard classified material IAW DoD 5220.22-M during the advanced phases of this contract.
PHASE I: Determine technical feasibility by investigating, evaluating (modeling and simulation), and identifying the most promising technical candidate approaches for AI-based real-time multimodal activity and evidence search, content tagging, sequencing, and discovery of information including the low signal-to-noise indicators that “connect the dots” with clues gathered from available networks and data sources, including cyber, financial, and social.
Perform trade-off studies among those approaches using actual datasets associated with events. Develop requirements, including scalability and multi-level security, for an open source trend and sentiment analysis framework. Recommend design, development, test, and prototyping requirements and a plan for Phase II. Deliverables include the final report evaluating the current state-of-the-art candidate approaches (pros and cons), test results and documentation, and recommended approach for Phase II.
Note 1: Phase-I will be UNCLASSIFIED and classified data is not required. For test and evaluation, a contractor needs to define the ground truth for a scenario and develop a storyboard to serve as an overarching scenario to guide the test and evaluation of this SBIR technology in a realistic context. Supporting datasets must have acceptable real- world data quality and complexity for the case studies to be considered rich in content. For example, image/video dataset of about 2,000 to 3,000 collected images for a case study can be considered content-rich.
Note 2: Contractors must provide appropriate dataset release authorization for use in their case studies, tests, and demonstrations, and certify that there are no legal or privacy issues, limitations, or restrictions with using the proposed data for this SBIR project.
PHASE II: Conduct proof-of-concept and prototype development for a scalable secure AI-platform incorporating the recommended candidate technologies from Phase I. Develop three plausible scenarios with data sources to support the prototype design. Develop performance metrics that will quantify the prototype’s capability for accurately measuring the correct direction and magnitude of processed sentiments and trends. Demonstrate scalability of the architecture and compatibility of the algorithms with cloud-based technologies. Verify and validate the performance and robustness of the system’s exploitation capacity. Develop detailed technology and transition plans for Phase III. Deliverables: System architecture and system interface requirements for mobile and stationary platforms, design documentation describing the techniques, prototype software, source code, user manuals, and a final report including test results.
Note: If Phase II prototyping, test, and validation require classified data, the proposal for Phase II work will be UNCLASSIFIED. If the selected Phase II contractor does not have the required certification for classified work, ONR or the related DON Program Office will work with the contractor to facilitate certification of related personnel and facility.
It is probable that the work under this effort will be classified under Phase II (see Description section for details).
PHASE III DUAL USE APPLICATIONS: Develop these capabilities to TRL-7 or 8 and integrate the technology into Minerva INP program suitable for ISR application supporting Naval Maritime Command and Control Operations Center and/or Marine Corps Information Operations Center. Once validated conceptually and technically, demonstrate dual use applications of this technology in civilian law enforcement, security services, and private security systems. In essence it enables rapid understanding of complex dynamic events and situations, and facilitates quick response by “connecting the dots” in an environment that involves a high volume of multimodal data types. It will have numerous knowledge management, behavior modeling and inference, situational awareness, and security applications in government, military, intelligence communities, law-enforcement, homeland security, and state and local governments to deal with asymmetric threats, deploying first responders, crisis management planning, and humanitarian aid response.
1. Pang, B. and Lee, L. “Opinion Mining and Sentiment Analysis.” Foundations and Trends in Information Retrieval, 2(1-2), 2008, pp. 1-135. http://www.cs.cornell.edu/home/llee/omsa/omsa.pdf
2. Dey, Lipika and Haque, SK Mirajul. “Opinion Mining From Noisy Text Data.” Proceedings of the second workshop on Analytics for noise unstructured text data, pp. 83.90. https://dl.acm.org/citation.cfm?id=1390763
3. Tan, Chenhao, Niculae, Vlad, Danescu-Niculescu-Mizil, and Lee, Lillian. “Winning Arguments: Interaction Dynamics and Persuasion Strategies in Good-faith Online Discussions.” Proceedings of the 25th International World Wide Web Conference (WWW'2016), April 11–15, 2016, Montréal, Québec, Canada.
4. Ma, Yiming, Liu, Bing, Wong, Ching Kian, Yu, P. S., and Lee, S. M. "Targeting the Right Students Using Data Mining." Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD-2000, Industry Track), Aug, 2000, Boston, USA. https://dl.acm.org/citation.cfm?doid=347090.347184
KEYWORDS: Artificial Intelligence; Latent-Risks; Trend; Sentiment; Machine Learning; Noisy Data; Behavior; Intent