Abstract: The discovery of quantum materials, from unconventional superconductors to topological and correlated systems, has traditionally relied on a slow and fragmented loop between theory, computation, and experiment. Serendipity and brute force trial-and-error have played an important role in many important discoveries. In this talk, I will describe how Periodic Labs is rethinking this loop by integrating experiments, large language models (LLMs), and physical theory into a unified discovery engine. Our approach aims to accelerate experimental iteration by automating characterization and analysis. I will discuss how LLMs can be used as reasoning interfaces over physical constraints, experimental data, and scientific literature. 

Ekin Dogus Cubuk is Co-Founder of Periodic Labs. Previously, he was a researcher at Google DeepMind where he works on deep learning and its applications to solid state physics and materials science. He received his Ph.D. from Harvard University where he studied the physics of disordered solids and battery materials using density functional theory and machine learning. After a brief postdoc at the Materials Science Department of Stanford University, he joined Google Brain in 2017. Since then, he has been studying the scaling and out-of-domain generalization properties of large neural networks, and their applications to materials discovery for applications including clean energy and information processing.