Adaptive human-in-the-loop AI solutions to rapidly changing problems
Areas of research
Malicious actor networks
Actors with malicious agendas organize in different network structures. Each has implications for groups or emergence coalitions of groups' capacity to use social media and the strategies and organizational structures they can build.
Natural Language Processing (NLP)
Our tools use NLP to extract data specific to our analysis from text and images, including video. In addition, we are building the capacity of the system to geo-locate images by, for example, matching them against known landmark databases or annotated image datasets.
Information Visualization (InfoVis) and Human-AI Interaction (HAI).
Social media is uniquely diverse, noisy, and multi-modal. Our domain experts have helped our AI specialists develop experiential machine learning models and on-the-fly visualizations to help users navigate this challenging data landscape.
Computer Vision (CV) and Multimodal Data Processing
Pioneering computer vision algorithms tailored to conflict-related images and data. This support includes, for example, geo-locating images by matching them against known landmark databases or annotated image datasets.
A global hub for integrated AI solutions
Our goal is to create a global, state-of-the-art platform to share human-in-the-loop tools, labels, models, and algorithms – both those created by our team and other innovators.