Yao Luo
Ph.D. Candidate · Applied Physics · Caltech
I am a Ph.D. candidate in Applied Physics at the California Institute of Technology, advised by Prof. Marco Bernardi. I received my B.S. in Physics from Nanjing University in 2020, graduating with honors.
My research sits at the intersection of computational physics, quantum many-body theory, and materials science. I develop first-principles methods to study quantum interactions in materials — including electron–phonon coupling, polaron physics, and anharmonic phonon interactions — and leverage machine learning and data-driven techniques to make these calculations more efficient and scalable. Recent highlights include a first-principles diagrammatic Monte Carlo approach for polarons and a tensor learning framework for compressing many-phonon interactions. I am a recipient of the 2024 Eddleman Graduate Fellowship.
My broader interest is to develop rigorous computational methods that unlock new physical observables and deepen our understanding of many-body quantum interactions in materials.
Polarons & Strong Coupling
Polarons — electrons dressed by lattice distortions — govern charge transport and optical properties in a wide class of materials, from complex oxides to halide perovskites. I develop first-principles diagrammatic Monte Carlo methods that sum Feynman diagrams to high order, enabling quantitatively accurate simulations of polaron formation, spectral functions, and mobility without empirical parameters [3]. Looking ahead, I aim to extend these approaches to nonequilibrium and finite-temperature regimes. This work opens pathways to predict and design materials with tailored polaronic properties. This research was featured on Caltech News ↗.
Data-Driven Compression of Quantum Interactions
First-principles calculations of quantum interactions — such as electron–phonon and phonon–phonon couplings — produce massive tensors that pose severe computational bottlenecks. I develop tensor learning and machine learning techniques to compress these interaction matrices by orders of magnitude while preserving full physical accuracy, making large-scale and high-order calculations tractable for realistic materials [2], [4]. The long-term vision is a general framework that autonomously learns compact representations of any quantum interaction, dramatically accelerating the discovery of new materials and physical phenomena. This research was featured on Caltech News: New AI Technique Unravels Quantum Atomic Vibrations ↗ and Speeding Up Calculations That Reveal How Electrons Interact ↗.
Teaching
Undergraduate Mentoring
- Dhruv Mangtani — Data-driven compression of phonon–phonon interactions
- Alex Roger — Spin relaxation in the polaron region using diagrammatic quantum Monte Carlo
I'm always happy to discuss research, collaborations, or questions about my work. Feel free to reach out by email.
Pasadena, CA 91125