Evaluating the impact of probabilistic and data-driven inference models on uncertainties of fiber-coupled NV-diamond thermometers

Published

May 5, 2025

Author(s)

Shraddha Rajpal, Zeeshan Ahmed, Tyrus Berry

Abstract

We conduct cw-Optically Detected Magnetic Resonance (ODMR) measurements using a fiber-coupled NV sensor to infer temperature. Our approach leverages a probabilistic feedforward inference model designed to maximize the likelihood of observed ODMR spectra through automatic differentiation. This model effectively utilizes the temperature dependence of spin Hamiltonian parameters to infer temperature from spectral features in the ODMR data. We achieve an accuracy of $\pm1 ^\circ$C across a temperature range of 243K to 323K. To benchmark our probabilistic model, we compare it with a non-parametric peak-finding technique and data-driven methodologies such as Principal Component Regression (PCR) and a 1D Convolutional Neural Network (CNN). We find that, within the temperature range of their training, data driven methods achieve a comparable accuracy of $\pm 1 ^\circ$C without incorporating expert-level understanding of the spectroscopic-temperature relationship. However, our results show that the probabilistic model outperforms both PCR and CNN when tested outside the training temperature range, indicating robustness and generalizability beyond the training set. In contrast, data-driven methods like PCR and CNN demonstrate significant challenges when tasked with extrapolating outside their training data range.

Citation

Optics Express

Pub Type

Journals

Download Paper

Local Download

Keywords

NV diamond, applied machine learning, uncertainty, temperature, deep learning

Citation

Rajpal, S. , Ahmed, Z. and Berry, T. (2025), Evaluating the impact of probabilistic and data-driven inference models on uncertainties of fiber-coupled NV-diamond thermometers, Optics Express, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=958356 (Accessed May 9, 2025)