Machine Learning and AI in Physics and Astronomy: Discovery, Data, and Education in the New Era

As machine learning becomes integral to scientific discovery, it challenges the traditional relationship between theory, experiment, and computation. In physics and astronomy, AI now complements first-principles modeling — not by replacing it, but by extending its reach. I will discuss how Physics-Informed Neural Networks (PINNs) bridge differential equations and data, enabling us to solve complex physical systems that were previously intractable. At the same time, the emergence of foundational models — architectures inspired by large language models (LLMs) — is reshaping the way we represent and reason about scientific data, from astronomical spectra to time-series observations.

In the second part of the talk, I’ll turn to education in the age of AI — exploring the key elements shaping a new teaching paradigm. Forces such as generative AI, the automation of analysis, and the availability of interactive intelligent agents are redefining what and how we teach. I’ll discuss how these pressures are shifting the emphasis from memorization and computation toward intuition, creativity, and critical reasoning, preparing students not just to use AI tools, but to think scientifically alongside them.

Speakers

• 

  Professor Pavlos Protopapas

  Scientific Program Director and Lecturer
  Harvard University