 |
Center for Nonlinear Studies |
Magnetic resonance pulse sequence techniques like the Carr–Purcell-Meiboom–Gill (CPMG) sequence are routinely used to dynamically decouple a qubit from an environmental field. This dynamical decoupling is routinely used to mitigate qubit decoherence due to dephasing (i.e. T2 errors). The optimal design of such pulse sequences relies on an understanding of the underlying spectral density of the noise source. However, gleaning information about these spectra from typical time-dynamics measurements on qubits is challenging. In this talk, we will discuss progress in noise learning by probing the noise spectra with various dynamical decoupling pulse sequences, with a focus on neural network and optimization strategies. Furthermore, we will discuss the novel numerical implementation of noise learning considering dynamical decoupling pulses with finite-width, improving upon noise learning using the traditional δ-function approximation. Our work enables robust, accurate noise learning from easily obtainable experimental measurements, and we provide an optimized, easy-to-use software package to implement this noise learning algorithm in Python.
Bio: Noah Huffman is a computational physicist with a background in experimental quantum optics. He received a BS in Physics and a BS in Business, Economics and Management from the California Institute of Technology in 2019, and his PhD in Physics from Stanford University in 2025. His current research interests are implementing algorithms for near-term quantum computers and the characterization and mitigation of errors to make them work. If he weren’t a physicist, he’d give stand-up comedy a try.