Researchers share their cutting-edge work in AI for national security

Earlier this year, DNN R&D sponsored the second workshop in a series on “Next-Generation AI for Proliferation Detection,” focused on domain-aware methods: computational techniques to combine domain information with data-driven AI models. While commercial AI is reaching great heights, the specific needs of nuclear nonproliferation require next-generation technologies uniquely suited to the field of national security.

The workshop: The workshop spanned four challenge areas specific to this mission:

• Complex and noisy environments.
• Sparse data and rare events.
• Robust deployment and decision support.
• Early proliferation detection and signature discovery.

The projects: Click here to read more about the following four projects presented by researchers who participated in the AI workshop:

• Global-scale cross-lingual proliferation expertise identification, monitoring, and forecasting.
• An artificial neural network system for special nuclear material detection in photon-based active interrogation scenarios.
• Applying domain-aware AI on the CBRN (chemical, biological, radiological, and nuclear weapons) battlefield.
• One-shot target detection via physics-informed training.