Robust Spin Relaxometry with Fast Adaptive Bayesian Estimation

Adaptive model‐based Magnetic Resonance

“…have been used to improve the precision of quantum sensors in estimating resonant spin frequencies in the presence of noise. [61][62][63][64] Adaptive schemes are also completely independent from other sources of acceleration, such as compressed sensing, MR fingerprinting, and parallel imaging, and can therefore be combined with them without detracting from their performance. Finally, unlike static sequences, adaptive acquisitions make almost no a priori assumptions on the range of the parameter being estimated; this is unlike static sequences, which must optimize their sequence parameters (e.g., TE) in anticipation of a specific range of tissue parameter (e.g., T 2 ).…”

“…Bayesian statistics provide an easy, straightforward framework for such updates. Indeed, adaptive Bayesian‐based estimations have been used for measurements involving nitrogen vacancy centers, where they have been used to improve the precision of quantum sensors in estimating resonant spin frequencies in the presence of noise 61–64 . Adaptive schemes are also completely independent from other sources of acceleration, such as compressed sensing, MR fingerprinting, and parallel imaging, and can therefore be combined with them without detracting from their performance.…”

Learning Quantum Systems

“…Experimentally, adaptive techniques have been employed, for example, to improve the sensitivity of optical phase sensing realizing a phase estimation protocol [224], or in quantum sensors based on the single spin associated with the nitrogen-vacancy (NV) centre in diamond, a system widely used for nanoscale magnetic mapping and magnetic resonance [151,152,[236][237][238][239]. Due to the time constraints, a key consideration for online sensing is the cost of the processing procedure: simplified nearoptimal methods might perform better than optimal but computationally-intensive ones.…”

“…(d) Different scenarios for optimised quantum control; time evolution, with control pulses as black arrows, measurements as orange arrows, agent in brown, and the quantum experiments' coherent evolution intervals in green. (e) Adaptive Bayesian inference and RL have been applied to or suggested for several experimental quantum systems, such as NV centres [151,152,[236][237][238][239], quantum dots [240,241], cavity-qubit systems [242,243] and multi-qubit systems with gates applied as actions and subject to projective measurements [244] .…”

Bayesian experimental design and parameter estimation for ultrafast spin dynamics

“…One promising avenue to streamline XPFS measurements is to couple data acquisition with Bayesian optimal experimental design (BOED) statistical methods [25][26][27][28] which have recently been harnessed in real experimental applications [29][30][31]. To achieve the most informed experimental design, physically realistic forward computations of system dynamics directly from model Hamiltonian are critically important.…”