Technical Proposal: Multimodal Analysis of Social Media to Detect Strategic Campaign Trends

Abstract

This effort will apply new techniques for data discovery and information triage to illuminate diagnostic patterns of deceptive and hostile information campaigns in order to develop perceptual methodologies for intuitive identification of social media patterns and trends insupport of US and NATO military-relevant strategic communication needs and requirements. We will apply these methods towa"rds the investigation of social media data~specifically content shared on the Twitter and Reddit platforms, including tweets/posts,"" account information, and relevant metadata~related to specific conversations, narratives, and/or themes that intersect with known o""r potential disinformation campaigns.The proposed research integrates audio, visual, and potentially other types of representation"s into this mixed method interpretive process~with the goal of 1) surfacing new insights about the spread of disinformation; 2) identifying new dimensions of signatures for understanding the spread of disinformation; and 3) determining specific opportunities for such new visual and audio representations to enhance information discovery by surfacing previously inaccessible insights. We aim to d"evelop clear, easy-to-understand perceptual models of raw datasets in orderto help operators and analysts intuitively and rapidly g"rasp changing trends in social media data~and with minimal cognitive load.

Document Details

Document Type
DoD Grant Award
Publication Date
Nov 06, 2017
Source ID
N000141812012

Entities

People

  • Jennifer Ziemke

Organizations

  • John Carroll University
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Strategic Security Studies