High-throughput ML/AI methods to use multiple data-streams from different diagnostics to characterize dynamic tokamak discharges

Modern magnetic fusion research involves high-resolution temporal and spatial diagnostics from multiple sensor arrays and provides opportunities to apply modern fusion-specific numerical linear algebra methods ({\it i}) to identify and optimize data reduction methods for real-time discharge control and ({\it ii}) to advance our understanding of fundamental behaviors of magnetically-confined plasma. This presentation uses measurements from recently expanded diagnostics on Columbia University's High Beta Tokamak-Extended Pulse (HBT-EP) that capture complex behaviors and records high-resolution, high-speed streams of magnetic, soft-x-ray, current, and optical data. The results of numerical analyses of these data streams from HBT-EP are examined, as well as how statistical methods such as the time-domain singular value decomposition and novel applications of methods from the field of ``randomized numerical linear algebra'' (rNLA) can be applied to fusion diagnostic data.