Where/When: Oak Ridge National Lab, June 8-12, 2026.
Purpose: Strengthen cross-cutting Data Science for Nuclear Nonproliferation by grounding students in mission relevant signals and constraints; training on modern & emerging AI (foundation models/agentic systems, RL, neuromorphic/edge, HPC); emphasizing rigorous evaluation (accuracy, low-FPR regimes, efficiency/latency, robustness); and building a pipeline from domain understanding → modeling → deployment.
Audience: NA-22 Consortium graduate and advanced undergraduate Fellows and Affiliates. Interested non-consortium students may also apply.
Format: ≈ 1/3 lectures, ≈ 1/3 lab scientist talks/tours, ≈ 1/3 hands-on.
Proposed Theme/Daily Topics (subject to possible change) Include:
Advanced modern and emerging, mission-aligned AI methods for nuclear nonproliferation, with a running competition on radiation detection and isotope identification using the RADAI dataset (Berkeley Data Cloud).
- NA-22 Mission, Signals, and the Data
- Foundation Models and Agentic Tooling
- Reinforcement Learning
- Neuromorphic/Edge AI + HPC Scaling and Optimization
- Final Push, Presentations, and Awards
This application’s deadline is February 15th, 2026. Please apply by then for first consideration.