Learning, Correcting, and Computing with Noisy Quantum Devices
Quantum computing is transitioning from small-scale demonstrations to devices large enough to tackle scientifically meaningful problems — but only if we can understand, mitigate, and ultimately correct the noise that limits them. In this talk, I will describe how three lines of research — quantum noise characterization, quantum error correction, and quantum algorithm design — are not independent challenges but deeply coupled facets of a single problem. Drawing on a body of work spanning scalable noise learning, noise-adapted error-correcting codes, and rigorous frameworks for learning from quantum experiments, I will show how these connections have recently enabled large-scale quantum simulations that access physics seemingly beyond the reach of current classical methods in three quantum computer architectures: quantum simulations of Fermi-Hubbard models in 56 trapped-ion and 72 superconducting qubits, as well as Floquet dynamics of disordered Heisenberg models on 100 superconducting qubits. Looking ahead, I will outline a research vision for noise-aware, error-correction-informed quantum algorithm co-design — an approach that tightly integrates device characterization with algorithm design and code selection, targeting early fault-tolerant simulation of strongly correlated quantum systems for practical quantum advantage.
Livestream the event on Zoom (Yale login required)