AI for Applied Physics Research

Graduate / Professional

Historically, the introduction of transformative tools—such as formal mathematics or the digital computer—fundamentally altered the methodology and scope of scientific inquiry. The computer, for instance, birthed simulations, computational optimization, and sophisticated curve-fitting. While now ubiquitous, these methods brought subtle but persistent challenges: numerical artifacts, convergence issues, and the potential loss of deep physical insight. Today, the rapid (and, frankly, surprising!) development of AI—from large language models to agentic systems—represents a similar “phase transition” in research. In disciplines like applied physics, where mathematics, code, and voluminous data intersect, this transition is both exciting and existential. This seminar examines how researchers are currently deploying these tools to advance the frontiers of their fields. We investigate parts of research where AI already or may soon exceed human performance, domains where humans are likely to retain a persistent advantage, identify common failure modes of AI-assisted research, and seek to critically develop “best practices” for rigorous, AI-augmented workflows. M 1:30pm-3:25pm