You are here

Characterization of Broadcast and Social Media for Indications and Warnings

Description:

TECHNOLOGY AREA(S): Info Systems 

OBJECTIVE: Develop capability to identify leading indicators of impending events/behaviors by identifying various (speech/text, video, image, cyber behaviors) leading indicators in broadcast and social media. 

DESCRIPTION: Much of the work in broadcast and social media has focused on “event analysis” – analysis of the “who’s”, the “what’s”, the “where’s”. For example, in text analytics, much of the focus is on entity extraction – extracting the “who’s” and “where’s” and “what’s” in order to derive situation understanding. However, there are distinct patterns in the nature of how broadcast and social media is used (tactics being employed such as “astroturfing”, diverting, etc., types of amplification or interest (e.g., retweets)). In addition, there are patterns in how people/organizations say or write or in visual content that give important clues about how they think about themselves and others and the “why’s” (why they may view another person or organization as a threat, why they are more likely to act out). These patterns in text, speech or videos and images (video memes or visual framing, visual symbols/cultural images) are often found prior (often weeks to months prior) to a behavior/event of interest; in many cases, before more obvious physical signals are present (coordination, reconnaissance, targeting, gathering resources/materials, etc.). Further, there are many signals in images and videos as well as in cyber behaviors. All of these are potentially informative for enabling forecasting of impending events. This SBIR would explore the use of various patterns in text/speech/images/videos/cyber behaviors in broadcast and/or social media in order to provide leading indicators of impending behaviors/events and make meaning regarding threat. This would have applicability to intelligence analysts, tipping and cueing of sensors, triggering adaptive planning, etc. for operations at tactical and operational levels. No government furnished materials, equipment, data, or facilities will be provided. 

PHASE I: Identify factors and/or features from either content (images, speech/text) and/or patterns of behavior in broadcast and social media which provide leading indicators of impending events/behaviors. Develop innovative techniques for extracting these features/factors and evaluate their performance for a single language and a single domain. The evaluation should evaluate both the semi-automated processing as well as the ability to support improved analysis. Due to the short time period of Phase I, it is preferable that currently available databases be used in the evaluation. 

PHASE II: Further develop the proposed techniques and evaluate their extensibility; that is, their performance for multiple events, languages and/or actors to show the generality of the techniques. The evaluations should follow the same format as described under the Phase I description but for the new events, new sources, and new domains. Any databases collected for development and/or evaluation should be delivered to the contract sponsor. 

PHASE III: Military applications include: Anticipatory ISR (e.g, Indications and Warnings (I&W), tipping and cueing), Humanitarian Relief, COA analysis, Stabilization and Reconstruction Operations (i.e., conflict resolution, negotiations). Commercial applications are similar to military applications but generally for different domains, such as: law enforcement, business (negotiations). 

REFERENCES: 

1: Fenstermacher, L. and Kuznar, L. 2016. Deciphering the "emic" perspective in data in order to assess threat. 2016. In Cohn, J., Schatz, S., and Freeman, H. Eds., Modeling Sociocultural Influences on Decision Making: Understanding Conflict, Enabling Stability. CRC Press LLC: Boca Raton, FL.

2: Hurley, C. M., Anker, A. E., Frank, M. G., Matsumoto, D. M., & Hwang, H. C. (2014). Background factors predicting accuracy and improvement in micro expression recognition. Motivation and Emotion, 38, 700714.

3: V.S. Subrahmanian and S. Kumar. Predicting Human Behavior: The Next Frontiers, Science Jan 2017.

4: Agarwal, S. Applying Social Media Intelligence for Predicting and Identifying On-line radicalization and Civil Unrest Oriented Threats. https://arxiv.org/pdf/1511.06858.

KEYWORDS: Leading Indicators, Forecasting, Indications And Warnings, Event Analysis, Text Analytics, Video Analysis, Cyber Behaviors, Integrative Complexity, Social Identity, Sentiment Analysis, Affect Analysis 

CONTACT(S): 

Laurie Fenstermacher (711 HPW/RHXM) 

(937) 255-0879 

laurie.fenstermacher@us.af.mil 

US Flag An Official Website of the United States Government