Google Scholar

Harrison Goldwyn's work at NLR focuses on developing interpretable, probabilistic artificial intelligence (AI) methods for scientific discovery, visualization, and autonomous experimentation. A central theme of his work is using Active Inference to model human decision-making, diagnose visualization-induced errors, and design trustworthy AI co-pilots for experimental systems. Goldwyn also investigates probabilistic reduced-order modeling methods, including PGQM, to improve basis selection, uncertainty-aware reconstruction, and computationally tractable analysis of complex simulation data.