-
Catching thermal avalanches in the disordered XXZ model
Authors:
Tomasz Szołdra,
Piotr Sierant,
Maciej Lewenstein,
Jakub Zakrzewski
Abstract:
We study the XXZ model with a random magnetic field in contact with a weakly disordered spin chain, acting as a finite thermal bath. We revise Fermi's golden rule description of the interaction between the thermal bath and the XXZ spin chain, contrasting it with a nonperturbative quantum avalanche scenario for the thermalization of the system. We employ two-point correlation functions to define th…
▽ More
We study the XXZ model with a random magnetic field in contact with a weakly disordered spin chain, acting as a finite thermal bath. We revise Fermi's golden rule description of the interaction between the thermal bath and the XXZ spin chain, contrasting it with a nonperturbative quantum avalanche scenario for the thermalization of the system. We employ two-point correlation functions to define the extent $ξ_d$ of the thermalized region next to the bath. Unbounded growth of $ξ_d$ proportional to the logarithm of time or faster is a signature of an avalanche. Such behavior signifies the thermalization of the system, as we confirm numerically for a generic initial state in the ergodic and critical regimes of the XXZ spin chain. In the many-body localized regime, a clear termination of avalanches is observed for specifically prepared initial states and, surprisingly, is not visible for generic initial product states. Additionally, we extract the localization length of the local integrals of motion and show that a bath made out of a weakly disordered XXZ chain has a similar effect on the system as a bath modeled by a Hamiltonian from a Gaussian orthogonal ensemble of random matrices. We also comment on the result of the earlier study (Phys. Rev. B 108, L020201 (2023)), arguing that the observed thermalization is due to external driving of the system and does not occur in the autonomous model. Our work reveals experimentally accessible signatures of quantum avalanches and identifies conditions under which termination of the avalanches may be observed.
△ Less
Submitted 9 April, 2024; v1 submitted 2 February, 2024;
originally announced February 2024.
-
Fidelity and Overlap of Neural Quantum States: Error Bounds on the Monte Carlo Estimator
Authors:
Tomasz Szołdra
Abstract:
Overlap between two neural quantum states can be computed through Monte Carlo sampling by evaluating the unnormalized probability amplitudes on a subset of basis configurations. Due to the presence of probability amplitude ratios in the estimator, which are possibly unbounded, convergence of this quantity is not immediately obvious. Our work provides a derivation of analytical error bounds on the…
▽ More
Overlap between two neural quantum states can be computed through Monte Carlo sampling by evaluating the unnormalized probability amplitudes on a subset of basis configurations. Due to the presence of probability amplitude ratios in the estimator, which are possibly unbounded, convergence of this quantity is not immediately obvious. Our work provides a derivation of analytical error bounds on the overlap in the Monte Carlo calculations as a function of their fidelity and the number of samples. Special case of normalized autoregressive neural quantum states is analyzed separately.
△ Less
Submitted 24 November, 2023;
originally announced November 2023.
-
Femtosecond pulse parameter estimation from photoelectron momenta using machine learning
Authors:
Tomasz Szołdra,
Marcelo F. Ciappina,
Nicholas Werby,
Philip H. Bucksbaum,
Maciej Lewenstein,
Jakub Zakrzewski,
Andrew S. Maxwell
Abstract:
Deep learning models have provided huge interpretation power for image-like data. Specifically, convolutional neural networks (CNNs) have demonstrated incredible acuity for tasks such as feature extraction or parameter estimation. Here we test CNNs on strong-field ionization photoelectron spectra, training on theoretical data sets to `invert' experimental data. Pulse characterization is used as a…
▽ More
Deep learning models have provided huge interpretation power for image-like data. Specifically, convolutional neural networks (CNNs) have demonstrated incredible acuity for tasks such as feature extraction or parameter estimation. Here we test CNNs on strong-field ionization photoelectron spectra, training on theoretical data sets to `invert' experimental data. Pulse characterization is used as a `testing ground', specifically we retrieve the laser intensity, where `traditional' measurements typically lead to 20% uncertainty. We report on crucial data augmentation techniques required to successfully train on theoretical data and return consistent results from experiments, including accounting for detector saturation. The same procedure can be repeated to apply CNNs in a range of scenarios for strong-field ionization. Using a predictive uncertainty estimation, reliable laser intensity uncertainties of a few percent can be extracted, which are consistently lower than those given by traditional techniques. Using interpretability methods can reveal parts of the distribution that are most sensitive to laser intensity, which can be directly associated with holographic interferences. The CNNs employed provide an accurate and convenient ways to extract parameters, and represent a novel interpretational tool for strong-field ionization spectra.
△ Less
Submitted 18 October, 2023; v1 submitted 24 March, 2023;
originally announced March 2023.
-
Tracking locality in time evolution of disordered systems
Authors:
Tomasz Szołdra,
Piotr Sierant,
Maciej Lewenstein,
Jakub Zakrzewski
Abstract:
Using local density correlation functions for a one-dimensional spin system, we introduce a correlation function difference (CFD) which compares correlations on a given site between a full system of size $L$ and its restriction to $\ell<L$ sites. We show that CFD provides useful information on transfer of information in quantum many-body systems by considering the examples of ergodic, Anderson, an…
▽ More
Using local density correlation functions for a one-dimensional spin system, we introduce a correlation function difference (CFD) which compares correlations on a given site between a full system of size $L$ and its restriction to $\ell<L$ sites. We show that CFD provides useful information on transfer of information in quantum many-body systems by considering the examples of ergodic, Anderson, and many-body localized regimes in disordered XXZ spin chain. In the ergodic phase, we find that the propagation of CFD is asymptotically faster than the spin transport but slower than the ballistic propagation implied by the Lieb-Robinson bound. In contrast, in the localized cases, we unravel an exponentially slow relaxation of CFD. Connections between CFD and other observables detecting non-local correlations in the system are discussed.
△ Less
Submitted 14 February, 2023; v1 submitted 21 November, 2022;
originally announced November 2022.
-
Unsupervised detection of decoupled subspaces: many-body scars and beyond
Authors:
Tomasz Szołdra,
Piotr Sierant,
Maciej Lewenstein,
Jakub Zakrzewski
Abstract:
Highly excited eigenstates of quantum many-body systems are typically featureless thermal states. Some systems, however, possess a small number of special, low-entanglement eigenstates known as quantum scars. We introduce a quantum-inspired machine learning platform based on a Quantum Variational Autoencoder (QVAE) that detects families of scar states in spectra of many-body systems. Unlike a clas…
▽ More
Highly excited eigenstates of quantum many-body systems are typically featureless thermal states. Some systems, however, possess a small number of special, low-entanglement eigenstates known as quantum scars. We introduce a quantum-inspired machine learning platform based on a Quantum Variational Autoencoder (QVAE) that detects families of scar states in spectra of many-body systems. Unlike a classical autoencoder, QVAE performs a parametrized unitary operation, allowing us to compress a single eigenstate into a smaller number of qubits. We demonstrate that the autoencoder trained on a scar state is able to detect the whole family of scar states sharing common features with the input state. We identify families of quantum many-body scars in the PXP model beyond the $\mathbb{Z}_2$ and $\mathbb{Z}_3$ families and find dynamically decoupled subspaces in the Hilbert space of disordered, interacting spin ladder model. The possibility of an automatic detection of subspaces of scar states opens new pathways in studies of models with a weak breakdown of ergodicity and fragmented Hilbert spaces.
△ Less
Submitted 5 June, 2022; v1 submitted 18 January, 2022;
originally announced January 2022.
-
Detecting ergodic bubbles at the crossover to many-body localization using neural networks
Authors:
Tomasz Szoldra,
Piotr Sierant,
Korbinian Kottmann,
Maciej Lewenstein,
Jakub Zakrzewski
Abstract:
The transition between ergodic and many-body localized phases is expected to occur via an avalanche mechanism, in which \emph{ergodic bubbles} that arise due to local fluctuations in system properties thermalize their surroundings leading to delocalization of the system, unless the disorder is sufficiently strong to stop this process. We propose an algorithm based on neural networks that allows to…
▽ More
The transition between ergodic and many-body localized phases is expected to occur via an avalanche mechanism, in which \emph{ergodic bubbles} that arise due to local fluctuations in system properties thermalize their surroundings leading to delocalization of the system, unless the disorder is sufficiently strong to stop this process. We propose an algorithm based on neural networks that allows to detect the ergodic bubbles using experimentally measurable two-site correlation functions. Investigating time evolution of the system, we observe a logarithmic in time growth of the ergodic bubbles in the MBL regime. The distribution of the size of ergodic bubbles converges during time evolution to an exponentially decaying distribution in the MBL regime, and a power-law distribution with a thermal peak in the critical regime, supporting thus the scenario of delocalization through the avalanche mechanism. Our algorithm permits to pin-point quantitative differences in time evolution of systems with random and quasiperiodic potentials, as well as to identify rare (Griffiths) events. Our results open new pathways in studies of the mechanisms of thermalization of disordered many-body systems and beyond.
△ Less
Submitted 18 October, 2021; v1 submitted 3 June, 2021;
originally announced June 2021.
-
Determination of Chern numbers with a phase retrieval algorithm
Authors:
Tomasz Szołdra,
Krzysztof Sacha,
Arkadiusz Kosior
Abstract:
Ultracold atoms in optical lattices form a clean quantum simulator platform which can be utilized to examine topological phenomena and test exotic topological materials. Here we propose an experimental scheme to measure the Chern numbers of two-dimensional multiband topological insulators with bosonic atoms. We show how to extract the topological invariants out of a sequence of time-of-flight imag…
▽ More
Ultracold atoms in optical lattices form a clean quantum simulator platform which can be utilized to examine topological phenomena and test exotic topological materials. Here we propose an experimental scheme to measure the Chern numbers of two-dimensional multiband topological insulators with bosonic atoms. We show how to extract the topological invariants out of a sequence of time-of-flight images by applying a phase retrieval algorithm to matter waves. We illustrate advantages of using bosonic atoms as well as efficiency and robustness of the method with two prominent examples: the Harper-Hofstadter model with an arbitrary commensurate magnetic flux and the Haldane model on a brick-wall lattice.
△ Less
Submitted 10 April, 2019; v1 submitted 13 December, 2018;
originally announced December 2018.
