ARLIS is excited to announce the selection of three INSURE partnership seedlings for funding to two members and one candidate member of the Intelligence and Security University Research Enterprise (INSURE) consortium. This announcement follows a competitive selection process for INSURE member-led and ARLIS-engaged research projects with high potential to support USG customer mission needs in future funded work and to position INSURE member faculty to more productively collaborate on ARLIS contracted work.
The call for proposals from ARLIS emphasized the need for projects that leverage cutting-edge technologies to enhance intelligence analysis, improve decision-making processes, and support operational planning. The team received 23 initial quad chart submissions from INSURE researchers and invited eight full proposals for further consideration based on their alignment with ARLIS customer demands. Following a thorough evaluation, the following three projects were selected for funding:
- Context-Aware Multimodal Information Retrieval Systems
Institution: University of Wisconsin
This initiative combines Large Multimodal Models (LMMs) with Retrieval-Augmented Generation (RAG) to deliver real-time, context-aware insights for national security applications. The system aims to enhance threat detection, situational awareness, operational planning, and overall decision-making processes. - AI Technology Evaluation Study via Testbed (AI TEST)
Institution: Arizona State University (ASU)
This project aims to improve understanding of human trust in AI by validating the “MASTOPIA” testbed, which measures context-dependent trust in AI-enabled systems. This testbed fills a gap in current test and evaluation, providing a cost-effective rubric that assesses the human-AI interdependencies that affect trust, a critical step in operationalizing AI systems for national defense. - Language-Driven Interaction with Geospatial Data
Institution: Washington University in St. Louis
This effort focuses on enabling non-expert users to interact with large-scale geospatial datasets through natural language interfaces, to segment and generate synthetic satellite imagery. This research aims to help more geospatial analysts become better at their jobs, making them more capable and efficient when working on national defense and intelligence tasks.
ARLIS is dedicated to fostering new collaborations that serve the national intelligence and security communities, aiming to bring value creation within the consortium. These funded projects represent a significant step forward in ARLIS’s mission areas and core competencies.
For further information on these projects and ARLIS's ongoing initiatives, please contact info@arlis.umd.edu.