Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2105.03127 · Physical Review A

Thermometry of one-dimensional Bose gases with neural networks

Frederik Møller, Thomas Schweigler, Mohammadamin Tajik, João Sabino, Federica Cataldini, Si-Cong Ji, Jörg Schmiedmayer

Published Oct 11, 2021• 10 citations
arXivHTMLPDF

Abstract

We design a neural network to extract and process features from absorption images taken of one-dimensional Bose gases in the quasicondensate regime. Specifically, the network is trained to predict both the temperature of single realizations of the system and the uncertainty thereof. For multiple realizations, the individual predictions can be combined in an estimate of the mean temperature, improving precision. We benchmark our model on both simulated and experimentally measured data and compare it to the established method of density ripple thermometry. We find the predictions of the two methods compatible, although the neural network reaches similar precision needing many fewer realizations, thus highlighting the efficiency gain achievable when incorporating neural networks into analysis of data from cold gas experiments. Further, we study feature maps to reveal which local features of the condensate are extracted by the network and how said features correlate with properties of the system. A similar analysis could be employed to uncover physical relations in more complex systems.

DOI
10.1103/physreva.104.043305

Related papers

Thermometry of simulated Bose–Einstein condensates using machine learning

J. Gwyn Griffiths, Steven A. Wrathmall, S. A. Gardiner

Abstract Precise determination of thermodynamic parameters in ultracold Bose gases remains challenging due to the destructive nature of conventional measurement techniques and inherent experimental uncertainties. We demonstrate a machine learning approach for rapid, minimally destructive estimation of the chemical potential and temperature from a single image of an in situ imaged density profiles of finite-temperature Bose gases. Our convolutional neural network is trained exclusively on quasi-two-dimensional ‘pancake’ condensates in harmonic trap configurations. It achieves parameter extraction within fractions of a second. The model also demonstrates some zero-shot generalisation across both trap geometry and thermalisation dynamics, successfully estimating the temperature (although not the chemical potential) for toroidally trapped condensates with errors of only a few nanokelvin despite no prior exposure to such geometries during training, and maintaining predictive accuracy during dynamic thermalisation processes after a relatively brief evolution without explicit training on non-equilibrium states. These results suggest that supervised learning can overcome traditional limitations in ultracold atom thermometry, with extension to broader geometric configurations, temperature ranges, and additional parameters potentially enabling comprehensive real-time analysis of quantum gas experiments. Such capabilities could significantly streamline experimental workflows whilst improving measurement precision across a range of quantum fluid systems.

arXiv:2506.16925•Machine Learning Science and Technology•Mar 9, 2026•0 citations

Momentum-Space Correlations of a One-Dimensional Bose Gas

Bess Fang, Aisling Johnson, Tommaso Roscilde +1 more

Analyzing the noise in the momentum profiles of single realizations of one-dimensional Bose gases, we present the experimental measurement of the full momentum-space density correlations ⟨δn_{p}δn_{p^{'}}⟩, which are related to the two-body momentum correlation function. Our data span the weakly interacting region of the phase diagram, going from the ideal Bose gas regime to the quasicondensate regime. We show experimentally that the bunching phenomenon, which manifests itself as super-Poissonian local fluctuations in momentum space, is present in all regimes. The quasicondensate regime is, however, characterized by the presence of negative correlations between different momenta, in contrast to the Bogolyubov theory for Bose condensates, predicting positive correlations between opposite momenta. Our data are in good agreement with ab initio calculations.

arXiv:1505.04675•Physical Review Letters•Feb 2, 2016•52 citations

Momentum distribution of one-dimensional Bose gases at the quasicondensation crossover: Theoretical and experimental investigation

T. Jacqmin, Bess Fang, Tarik Berrada +2 more

We investigate the momentum distribution of weakly interacting one-dimensional (1D) Bose gases at thermal equilibrium both experimentally and theoretically. Momentum distribution of 1D Bose gases is measured using a focusing technique, whose resolution we improve via a guiding scheme. The momentum distribution compares very well with quantum Monte Carlo calculations for the Lieb-Liniger model at finite temperature, allowing for an accurate thermometry of the gas that agrees with (and improves upon) the thermometry based on in situ density fluctuation measurements. The quasicondensation crossover is investigated via two different experimental parameter sets, corresponding to the two different sides of the crossover. Classical field theory is expected to correctly describe the quasicondensation crossover of weakly interacting gases. We derive the condition of validity of the classical field theory, and find that, in typical experiments, interactions are too strong for this theory to be accurate. This is confirmed by a comparison between the classical field predictions and the numerically exact quantum Monte Carlo calculations.

arXiv:1207.2855•Physical Review A•Oct 25, 2012•58 citations

Neural-network quantum states for ultra-cold Fermi gases

Jane Kim, Gabriel Pescia, Bryce Fore +5 more

Abstract Ultra-cold Fermi gases exhibit a rich array of quantum mechanical properties, including the transition from a fermionic superfluid Bardeen-Cooper-Schrieffer (BCS) state to a bosonic superfluid Bose-Einstein condensate (BEC). While these properties can be precisely probed experimentally, accurately describing them poses significant theoretical challenges due to strong pairing correlations and the non-perturbative nature of particle interactions. In this work, we introduce a Pfaffian-Jastrow neural-network quantum state featuring a message-passing architecture to efficiently capture pairing and backflow correlations. We benchmark our approach on existing Slater-Jastrow frameworks and state-of-the-art diffusion Monte Carlo methods, demonstrating a performance advantage and the scalability of our scheme. We show that transfer learning stabilizes the training process in the presence of strong, short-ranged interactions, and allows for an effective exploration of the BCS-BEC crossover region. Our findings highlight the potential of neural-network quantum states as a promising strategy for investigating ultra-cold Fermi gases.

Communications Physics•May 8, 2024•37 citations

Fluctuation Thermometry of an Atom-Resolved Quantum Gas: Beyond the Fluctuation-Dissipation Theorem

Maxime Dixmerias, Joris Verstraten, Cyprien Daix +3 more

Thermometry is essential for studying many-body physics with ultracold atoms. Accurately measuring low temperatures in these systems, however, remains a significant challenge due to the absence of a universal thermometer. Most widely applicable methods, such as fitting of in situ density profiles, are limited by the requirement of global thermal equilibrium and inapplicability to homogeneous systems. In this Letter, we introduce a novel in situ thermometry for quantum gases, leveraging single-atom-resolved measurements via quantum gas microscopy, and demonstrate it on an ideal Fermi gas. By analyzing number fluctuations in probe volumes containing as little as one atom on average, we extract both global and local temperatures over a broad dynamic range. Our method does not rely on the fluctuation-dissipation theorem and is based instead on the exact relationship between number fluctuations and density-density correlations. In the low-temperature regime, it allows us to observe significant deviations from fluctuation-dissipation predictions, uncovering subextensive fluctuations. Our method is applicable to systems with arbitrary trapping potentials, requiring neither precise trap calibration nor global thermal equilibrium. This thermometer for quantum gases overcomes key limitations of existing techniques, paving the way for more accurate and versatile temperature measurements in ultracold quantum systems.

arXiv:2502.05132•Physical Review Letters•Jun 12, 2026•0 citations

Mode-resolved thermometry of a trapped ion with deep learning

Yi Tao, Ting Chen, Hongyang Wang +5 more

In a trapped-ion system, accurate thermometry of ions is crucial for precisely evaluating the system state and performing quantum operations. However, when the motional state of a single ion is far away from the ground state, the spatial dimensionality of the phonon state sharply increases. Then it is difficult to realize accurate and mode-resolved thermometry with existing methods. In this work, we apply deep learning to the thermometry of a trapped ion and propose an efficient and mode-resolved method for accurately estimating large mean phonon numbers. We have also conducted experimental verification based on a self-developed surface trap. The result has shown the significant accuracy and efficiency of the method for thermometry of a single ion with large mean phonon numbers. The mode-resolved feature of our method makes it better applied to the characterization of system parameters, such as evaluating cooling effectiveness and analyzing surface trap noise. Our trained neural network model can be easily extended to other experimental setups and parameter ranges without modification of the experimental setups.

arXiv:2402.19022•Physical Review A•Jun 25, 2024•5 citations
We design a neural network to extract and process features from absorption images taken of one-dimensional Bose gases in the quasicondensate regime. Specifically, the network is trained to predict both the temperature of single realizations of the system and the uncertainty thereof. For multiple realizations, the individual predictions can be combined in an estimate of the mean temperature, improving precision. We benchmark our model on both simulated and experimentally measured data and compare it to the established method of density ripple thermometry. We find the predictions of the two methods compatible, although the neural network reaches similar precision needing many fewer realizations, thus highlighting the efficiency gain achievable when incorporating neural networks into analysis of data from cold gas experiments. Further, we study feature maps to reveal which local features of the condensate are extracted by the network and how said features correlate with properties of the system. A similar analysis could be employed to uncover physical relations in more complex systems.
Abstract Precise determination of thermodynamic parameters in ultracold Bose gases remains challenging due to the destructive nature of conventional measurement techniques and inherent experimental uncertainties. We demonstrate a machine learning approach for rapid, minimally destructive estimation of the chemical potential and temperature from a single image of an in situ imaged density profiles of finite-temperature Bose gases. Our convolutional neural network is trained exclusively on quasi-two-dimensional ‘pancake’ condensates in harmonic trap configurations. It achieves parameter extraction within fractions of a second. The model also demonstrates some zero-shot generalisation across both trap geometry and thermalisation dynamics, successfully estimating the temperature (although not the chemical potential) for toroidally trapped condensates with errors of only a few nanokelvin despite no prior exposure to such geometries during training, and maintaining predictive accuracy during dynamic thermalisation processes after a relatively brief evolution without explicit training on non-equilibrium states. These results suggest that supervised learning can overcome traditional limitations in ultracold atom thermometry, with extension to broader geometric configurations, temperature ranges, and additional parameters potentially enabling comprehensive real-time analysis of quantum gas experiments. Such capabilities could significantly streamline experimental workflows whilst improving measurement precision across a range of quantum fluid systems.
Analyzing the noise in the momentum profiles of single realizations of one-dimensional Bose gases, we present the experimental measurement of the full momentum-space density correlations ⟨δn_{p}δn_{p^{'}}⟩, which are related to the two-body momentum correlation function. Our data span the weakly interacting region of the phase diagram, going from the ideal Bose gas regime to the quasicondensate regime. We show experimentally that the bunching phenomenon, which manifests itself as super-Poissonian local fluctuations in momentum space, is present in all regimes. The quasicondensate regime is, however, characterized by the presence of negative correlations between different momenta, in contrast to the Bogolyubov theory for Bose condensates, predicting positive correlations between opposite momenta. Our data are in good agreement with ab initio calculations.
We investigate the momentum distribution of weakly interacting one-dimensional (1D) Bose gases at thermal equilibrium both experimentally and theoretically. Momentum distribution of 1D Bose gases is measured using a focusing technique, whose resolution we improve via a guiding scheme. The momentum distribution compares very well with quantum Monte Carlo calculations for the Lieb-Liniger model at finite temperature, allowing for an accurate thermometry of the gas that agrees with (and improves upon) the thermometry based on in situ density fluctuation measurements. The quasicondensation crossover is investigated via two different experimental parameter sets, corresponding to the two different sides of the crossover. Classical field theory is expected to correctly describe the quasicondensation crossover of weakly interacting gases. We derive the condition of validity of the classical field theory, and find that, in typical experiments, interactions are too strong for this theory to be accurate. This is confirmed by a comparison between the classical field predictions and the numerically exact quantum Monte Carlo calculations.
Abstract Ultra-cold Fermi gases exhibit a rich array of quantum mechanical properties, including the transition from a fermionic superfluid Bardeen-Cooper-Schrieffer (BCS) state to a bosonic superfluid Bose-Einstein condensate (BEC). While these properties can be precisely probed experimentally, accurately describing them poses significant theoretical challenges due to strong pairing correlations and the non-perturbative nature of particle interactions. In this work, we introduce a Pfaffian-Jastrow neural-network quantum state featuring a message-passing architecture to efficiently capture pairing and backflow correlations. We benchmark our approach on existing Slater-Jastrow frameworks and state-of-the-art diffusion Monte Carlo methods, demonstrating a performance advantage and the scalability of our scheme. We show that transfer learning stabilizes the training process in the presence of strong, short-ranged interactions, and allows for an effective exploration of the BCS-BEC crossover region. Our findings highlight the potential of neural-network quantum states as a promising strategy for investigating ultra-cold Fermi gases.
Thermometry is essential for studying many-body physics with ultracold atoms. Accurately measuring low temperatures in these systems, however, remains a significant challenge due to the absence of a universal thermometer. Most widely applicable methods, such as fitting of in situ density profiles, are limited by the requirement of global thermal equilibrium and inapplicability to homogeneous systems. In this Letter, we introduce a novel in situ thermometry for quantum gases, leveraging single-atom-resolved measurements via quantum gas microscopy, and demonstrate it on an ideal Fermi gas. By analyzing number fluctuations in probe volumes containing as little as one atom on average, we extract both global and local temperatures over a broad dynamic range. Our method does not rely on the fluctuation-dissipation theorem and is based instead on the exact relationship between number fluctuations and density-density correlations. In the low-temperature regime, it allows us to observe significant deviations from fluctuation-dissipation predictions, uncovering subextensive fluctuations. Our method is applicable to systems with arbitrary trapping potentials, requiring neither precise trap calibration nor global thermal equilibrium. This thermometer for quantum gases overcomes key limitations of existing techniques, paving the way for more accurate and versatile temperature measurements in ultracold quantum systems.
In a trapped-ion system, accurate thermometry of ions is crucial for precisely evaluating the system state and performing quantum operations. However, when the motional state of a single ion is far away from the ground state, the spatial dimensionality of the phonon state sharply increases. Then it is difficult to realize accurate and mode-resolved thermometry with existing methods. In this work, we apply deep learning to the thermometry of a trapped ion and propose an efficient and mode-resolved method for accurately estimating large mean phonon numbers. We have also conducted experimental verification based on a self-developed surface trap. The result has shown the significant accuracy and efficiency of the method for thermometry of a single ion with large mean phonon numbers. The mode-resolved feature of our method makes it better applied to the characterization of system parameters, such as evaluating cooling effectiveness and analyzing surface trap noise. Our trained neural network model can be easily extended to other experimental setups and parameter ranges without modification of the experimental setups.