Intent driven adversarial modeling is the corner-stone of a military decision-making process. To 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, we are developing WINSTON (Warnings and Indicators from Net-like distributed Signatures of Threats in Open communications and News media) -- a novel computational framework for using the open source data (e.g., internet-based social media and news media) about the Gray Zones of interest, in order to stage: (1) semi-automated profiling of relevant actors (organizations, groups, key individuals), their threat intentions and actions/tactics, as well as threat types; (2) semi-automatic extraction, association, and analysis of data from multiple channels and modalities to detect, monitor, and focus on the emergent and existing threats; and (3) detecting the weeks-to-months-prior leading indicators (in text/speech, images/videos, cyber behaviors) and signature patterns of hostility (including discourse content patterns and attitude cues; capability and mental state changes; distributed event patterns; multi-actor interaction signatures) in social media and news broadcasts.Gray Zones,Open Source Intelligence,Evidence-based explanations,Leading indicators of hostility,Adversarial actors,adversarial intent,Signature patterns
