Cultural Events Calendar

Over the last half decade, the capabilities of large language models (like ChatGPT and Gemini) have leapt from babbling preschoolers to International Math Olympiad gold medallists, and now beyond. This talk reviews recent progress in training Artificial Intelligences to do science and reasoning, and speculates as to what it will mean for the future of science if these trends continue.

The lecture will be followed by a discussion panel featuring leading artificial intelligence experts, offering multiple perspectives on the opportunities and challenges shaping AI today.

Artificial Intelligence (AI) I is a transformative technology that is quickly raising the ambitions of scientists. Last year’s Nobel prizes in physics and chemistry highlighted the significance of these developments. However, the capabilities that AI enables and the ways those capabilities fit into the scientific method vary significantly. The incorporation of AI into scientific workflows raises important questions. For example, how do we maintain scientific rigor when incorporating AI components that are approximate or may ‘hallucinate’? These emerging patterns also are giving rise to a new set of questions in the philosophy of science. What is the role of interpretability, causality, prediction, hypothesis generation, etc.? What is the role of human understanding?

A hundred years ago, we did not know what atoms were really made of, or why they emitted light only of very specific colors. That puzzle led to the invention of quantum mechanics. The physicists of the 1920s were trying to understand nature, yet from that work came lasers, modern electronics, GPS, and a long list of everyday technologies.

A century later, we have learned to do quantum mechanics in the laboratory. We can cool, trap, and control individual atoms, and use them to build quantum atomic clocks so precise that they would not lose even a fraction of a second over the entire time since the Big Bang. These clocks are so sensitive that lifting one by millimeters measurably changes its tick rate due to changes in in Earth’s gravity.

Meanwhile, we face a new and deeper puzzle: we do not know what most of the universe is made of. Decades of observations point to dark matter and dark energy that dominate the cosmos, yet their nature remains unknown. I will describe how quantum clocks work can act as new observatories for this invisible universe, searching for subtle drifts in their ticking that could signal dark matter, and testing gravity on Earth and, in the future, in space.

Our best theoretical models struggle to explain how the early Universe built so many giants so quickly, or why they stopped forming stars and retired so early. In this talk I’ll describe how my team is hunting down these cosmic overachievers and what their existence means for our understanding of how the Universe evolved.

In science, random fluctuations (e.g., noise) are typically regarded as a nuisance to be minimized or avoided if possible. Yet in many important scenarios, valuable information can be gleaned from their careful study. In this talk, I will discuss three such examples, from different disciplines.

In science, random fluctuations (e.g., noise) are typically regarded as a nuisance to be minimized or avoided if possible. Yet in many important scenarios, valuable information can be gleaned from their careful study. In this talk, I will discuss three such examples, from different disciplines.