arXiv subjects

Tomi Ohtsuki

Publications and source records attributed to Tomi Ohtsuki.

At least 19 recordsLinked to original sources

arXiv2026

Reservoir computing by thin film embedded with magnetic impurities

The reservoir computing based on the thin film embedded with magnetic impurities in the presence of the long-range (the dipole-dipole) interaction is numerically investigated. We simulated the magnetization dynamics by taking into account the dipole-dipole interaction and performed the handwritten-digit recognition task. Although the training data is prepared by taking spatial average in the sample, the high classification accuracy is achieved. Our result demonstrates that the long range interaction effectively encodes the complex spatial input pattern into the time domain, even when only a spatially averaged output is accessible. The proposed system paves the way for easily realizable magnetic reservoir computing.

cond-mat.dis-nn↗

arXiv2025

Adaptive Interpolating Quantum Transform: A Quantum-Native Framework for Efficient Transform Learning

Machine learning on quantum computers has attracted attention for its potential to deliver computational speedups in different tasks. However, deep variational quantum circuits require a large number of trainable parameters that grows with both qubit count and circuit depth, often rendering training infeasible. In this study, we introduce the Adaptive Interpolating Quantum Transform (AIQT), a quantum-native framework for flexible and efficient learning. AIQT defines a trainable unitary that interpolates between quantum transforms, such as the Hadamard and quantum Fourier transforms. This approach enables expressive quantum state manipulation while controlling parameter overhead. It also allows AIQT to inherit any quantum advantages present in its constituent transforms. Our results show that AIQT achieves high performance with minimal parameter count, offering a scalable and interpretable alternative to deep variational circuits.

quant-ph↗

arXiv2025

Backsolution: A Framework for Solving Inverse Problems via Automatic Differentiation

We present a simple yet powerful framework for solving inverse problems by leveraging automatic differentiation. Our method is broadly applicable whenever a smooth cost function can be defined near the true solution, and a numerical simulator is available. As a concrete example, we demonstrate that our method can accurately reconstruct the spatial profiles in a conductor from magnetotransport measurements. Even if the given data are insufficient to uniquely determine the profiles, the same framework enables effective reverse modeling. This method is general, flexible, and readily adaptable to a broad class of inverse problems across condensed matter physics and beyond.

cond-mat.dis-nn↗

arXiv2025

Universal Stochastic Equations of Monitored Quantum Dynamics

We investigate the monitored quantum dynamics of Gaussian mixed states and derive the universal Fokker-Planck equations that govern the stochastic time evolution of entire density-matrix spectra, obtaining their exact solutions. From these equations, we reveal an even-odd effect in purification dynamics: whereas entropy exhibits exponential decay for an even number $N$ of complex fermions, algebraic decay with divergent purification time occurs for odd $N$ as a manifestation of dynamical criticality. Additionally, we identify the universal fluctuations of entropy in the chaotic regime, serving as a non-unitary counterpart of the universal conductance fluctuations in mesoscopic electronic transport phenomena. Furthermore, we elucidate and classify the universality classes of non-unitary quantum dynamics based on fundamental symmetry. We also validate the universality of these analytical results through extensive numerical simulations across different types of models.

cond-mat.stat-mech↗

arXiv2025

Forecasting long-time dynamics in quantum many-body systems by dynamic mode decomposition

Reliable numerical computation of quantum dynamics is a fundamental challenge when the long-ranged quantum entanglement plays essential roles as in the cases governed by quantum criticality in strongly correlated systems. Here we apply a method that utilizes reliable short-time data of physical quantities to accurately forecast long-time behavior of the strongly entangled systems. We straightforwardly employ the simple dynamic mode decomposition (DMD), which is commonly used in fluid dynamics. Despite the simplicity of the method, the effectiveness and applicability of the DMD in quantum many-body systems such as the Ising model in the transverse field at the critical point are demonstrated, even when the time evolution at long time exhibits complicated features such as a volume-law entanglement entropy and consequential power-law decays of correlations characteristic of systems with long-ranged quantum entanglements unlike fluid dynamics. The present method, though simple, enables accurate forecasts amazingly at time as long as nearly an order of magnitude longer than that of the short-time training data. Effects of noise on the accuracy of the forecast are also investigated, because they are important especially when dealing with the experimental data. We find that a few percentages of noise do not affect the prediction accuracy destructively.

quant-ph↗

arXiv2024

Negative longitudinal resistance of monolayer graphene in the quantum Hall regime

In the quantum Hall regime the charge current is carried by ideal one-dimensional edge channels where the backscattering is prohibited by topology. This results in the constant potential along the edge of the Hall bar leading to zero 4-terminal longitudinal resistance r_xx. Finite scattering between the counter-propagating edge states, when the topological protection is broken, commonly results in r_xx > 0. However, a local disorder, if allowing intersection of the edge states, can result in a counter-intuitive scenario when r_xx<0. In this work we report the observation and a systematic study of such unconventional negative longitudinal resistance seen in an encapsulated monolayer graphene Hall bar device measured in the quantum Hall regime. We supplement our findings with the numerical calculations which allow us to outline the conditions necessary for the appearance of negative r_xx and to exclude the macroscopic disorder (contamination bubble) as the main origin of it.

cond-mat.mes-hall↗

arXiv2023

Channel Attention for Quantum Convolutional Neural Networks

Quantum convolutional neural networks (QCNNs) have gathered attention as one of the most promising algorithms for quantum machine learning. Reduction in the cost of training as well as improvement in performance is required for practical implementation of these models. In this study, we propose a channel attention mechanism for QCNNs and show the effectiveness of this approach for quantum phase classification problems. Our attention mechanism creates multiple channels of output state based on measurement of quantum bits. This simple approach improves the performance of QCNNs and outperforms a conventional approach using feedforward neural networks as the additional post-processing.

quant-ph↗

arXiv2023

Singular-Value Statistics of Non-Hermitian Random Matrices and Open Quantum Systems

The spectral statistics of non-Hermitian random matrices are of importance as a diagnostic tool for chaotic behavior in open quantum systems. Here, we investigate the statistical properties of singular values in non-Hermitian random matrices as an effective measure of quantifying dissipative quantum chaos. By means of Hermitization, we reveal the unique characteristics of the singular-value statistics that distinguish them from the complex-eigenvalue statistics, and establish the comprehensive classification of the singular-value statistics for all the 38-fold symmetry classes of non-Hermitian random matrices. We also analytically derive the singular-value statistics of small random matrices, which well describe those of large random matrices in the similar spirit to the Wigner surmise. Furthermore, we demonstrate that singular values of open quantum many-body systems follow the random-matrix statistics, thereby identifying chaos and nonintegrability in open quantum systems. Our work elucidates that the singular-value statistics serve as a clear indicator of symmetry and lay a foundation for statistical physics of open quantum systems.

cond-mat.mes-hall↗

arXiv2023

Anisotropic Topological Anderson Transitions in Chiral Symmetry Classes

We study quantum phase transitions of three-dimensional disordered systems in the chiral classes (AIII and BDI) with and without weak topological indices. We show that the systems with a nontrivial weak topological index universally exhibit an emergent thermodynamic phase where wave functions are delocalized along one spatial direction but exponentially localized in the other two spatial directions, which we call the quasi-localized phase. Our extensive numerical study clarifies that the critical exponent of the Anderson transition between the metallic and quasi-localized phases, as well as that between the quasi-localized and localized phases, are different from that with no weak topological index, signaling the new universality classes induced by topology. The quasi-localized phase and concomitant topological Anderson transition manifest themselves in the anisotropic transport phenomena of disordered weak topological insulators and nodal-line semimetals, which exhibit the metallic behavior in one direction but the insulating behavior in the other directions.

cond-mat.dis-nn↗

arXiv2023

Critical behavior of Anderson transitions in higher dimensional Bogoliubov-de Gennes symmetry classes

Disorder is ubiquitous in solid-state systems, and its crucial influence on transport properties was revealed by the discovery of Anderson localization. Generally speaking, all bulk states will be exponentially localized in the strong disorder limit, but whether an Anderson transition takes place depends on the dimension and symmetries of the system. The scaling theory and symmetry classes are at the heart of the study of the Anderson transition, and the critical exponent $ν$ characterizing the power-law divergence of localization length is of particular interest. In contrast with the well-established lower critical dimension $d_l=2$ of the Anderson transition, the upper critical dimension $d_u$, above which the disordered system can be described by mean-field theory, remains uncertain, and precise numerical evaluations of the critical exponent in higher dimensions are needed. In this study, we apply Borel-Padé resummation method to the known perturbative results of the non-linear sigma model (NL$σ$M) to estimate the critical exponents of the Boguliubov-de Gennes (BdG) classes. We also report numerical simulations of class DIII in 3D, and classes C and CI in 4D, and compare the results of the resummation method with these and previously published work. Our results may be experimentally tested in realizations of quantum kicked rotor models in atomic-optic systems, where the critical behavior of dynamical localization in higher dimensions can be measured.

cond-mat.dis-nn↗

arXiv2023

Machine learning wave functions to identify fractal phases

We demonstrate that an image recognition algorithm based on a convolutional neural network provides a powerful procedure to differentiate between ergodic, non-ergodic extended (fractal) and localized phases in various systems: single-particle models, including random-matrix and random-graph models, and many-body quantum systems. The network can be successfully trained on a small data set of only 500 wave functions (images) per class for a single model. The trained network can then be used to classify phases in the other models and is thus very efficient. We discuss the strengths and limitations of the approach.

cond-mat.dis-nn↗

arXiv2023

Irrelevant corrections at the quantum Hall transition

The quantum Hall effect is one of the most extensively studied topological effects in solid state physics. The transitions between different quantum Hall states exhibit critical phenomena described by universal critical exponents. Numerous numerical finite size scaling studies have focused on the critical exponent $ν$ for the correlation length. In such studies it is important to take proper account of irrelevant corrections to scaling. Recently, Dresselhaus et al. [Phys. Rev. Lett. {\bf 129}, 026801(2022)] proposed a new scaling ansatz, which they applied to the two terminal conductance of the Chalker-Coddington model. In this paper their proposal is applied to our previously reported data for the Lyapunov exponents of that model using both polynomial fitting and Gaussian process fitting.

cond-mat.dis-nn↗

arXiv2022

Level statistics of real eigenvalues in non-Hermitian systems

Symmetries associated with complex conjugation and Hermitian conjugation, such as time-reversal symmetry and pseudo-Hermiticity, have great impact on eigenvalue spectra of non-Hermitian random matrices. Here, we show that time-reversal symmetry and pseudo-Hermiticity lead to universal level statistics of non-Hermitian random matrices on and around the real axis. From the extensive numerical calculations of large random matrices, we obtain the five universal level-spacing and level-spacing-ratio distributions of real eigenvalues, each of which is unique to the symmetry class. Furthermore, we analyse spacings of real eigenvalues in physical models, such as bosonic many-body systems and free fermionic systems with disorder and dissipation. We clarify that the level spacings in ergodic (metallic) phases are described by the universal distributions of non-Hermitian random matrices in the same symmetry classes, while the level spacings in many-body localized and Anderson localized phases show the Poisson statistics. We also find that the number of real eigenvalues shows distinct scalings in the ergodic and localized phases in these symmetry classes. These results serve as effective tools for detecting quantum chaos, many-body localization, and real-complex transitions in non-Hermitian systems with symmetries.

cond-mat.dis-nn↗

arXiv2022

Universality classes of the Anderson transitions driven by quasiperiodic potential in the three-dimensional Wigner-Dyson symmetry classes

Quasiperiodic system is an intermediate state between periodic and disordered systems with unique delocalization-localization transition driven by the quasiperiodic potential (QP). One of the intriguing questions is whether the universality class of the Anderson transition (AT) driven by QP is similar to that of the AT driven by the random potential in the same symmetry class. Here, we study the critical behavior of the ATs driven by QP in the three-dimensional (3D) Anderson model, Peierls phase model, and Ando model, which belong to the Wigner-Dyson symmetry classes. The localization length and two-terminal conductance have been calculated by the transfer matrix method, and we argue that their error estimations in statistics suffer from the correlation of QP. With the correlation under control, the critical exponents $ν$ of the ATs driven by QP are estimated by the finite size scaling analysis of conductance, which are consistent with $ν$'s of the ATs driven by the random potential. Moreover, the critical conductance distribution and the level spacing ratio distribution have been studied. We also find that a convolutional neural network trained by the localized/delocalized wavefunctions in a disordered system predicts the localized/delocalized wavefunctions in quasiperiodic systems. Our numerical results strongly support that the universality classes of the ATs driven by QP and random potential are similar in the 3D Wigner-Dyson symmetry classes.

cond-mat.dis-nn↗

arXiv2022

Quantum phase transition between hyperuniform density distributions

We study an electron distribution under a quasiperiodic potential in light of hyperuniformity, aiming to establish a classification and analysis method for aperiodic but orderly density distributions realized in, e.g., quasicrystals. Using the Aubry-Andre-Harper model, we first reveal that the electron-charge distribution changes its character as the increased quasiperiodic potential alters the eigenstates from extended to localized ones. While these changes of the charge distribution are characterized by neither multifractality nor translational-symmetry breaking, they are characterized by hyperuniformity class and its order metric. We find a nontrivial relationship between the density of states at the Fermi level, a charge-distribution histogram, and the hyperuniformity class. The change to a different hyperuniformity class occurs as a first-order phase transition except for an electron-hole symmetric point, where the transition is of the third order. Moreover, we generalize the hyperuniformity order metric to a function, to capture more detailed features of the density distribution, in some analogy with a generalization of the fractal dimension to a multifractal one.

cond-mat.str-el↗

arXiv2022

Deciphering quantum fingerprints in electric conductance

When the electric conductance of a nano-sized metal is measured at low temperatures, it often exhibits complex but reproducible patterns as a function of external magnetic fields, called quantum fingerprints in electric conductance. Such complex patterns are due to quantum-mechanical interference of conduction electrons; when thermal disturbance is feeble and coherence of the electrons extends all over the sample, the quantum interference pattern reflects microscopic structures, such as crystalline defects and the shape of the sample, giving rise to complicated interference. Although the interference pattern carries such microscopic information, it looks so random that it has not been analysed. Here we show that machine learning allows us to decipher quantum fingerprints; fingerprint patterns in magneto-conductance are shown to be transcribed into spatial images of electron wave function intensities (WIs) in a sample by using generative machine learning. The output WIs reveal quantum interference states of conduction electrons, as well as sample shapes. The present result augments the human ability to identify quantum states, and it should allow microscopy of quantum nanostructures in materials by making use of quantum fingerprints.

cond-mat.mes-hall↗

arXiv2022

Unifying the Anderson Transitions in Hermitian and Non-Hermitian Systems

Non-Hermiticity enriches the 10-fold Altland-Zirnbauer symmetry class into the 38-fold symmetry class, where critical behavior of the Anderson transitions (ATs) has been extensively studied recently. Here, we propose a correspondence of the universality classes of the ATs between Hermitian and non-Hermitian systems. We illustrate that the critical exponents of the length scale in non-Hermitian systems coincide with the critical exponents in the corresponding Hermitian systems with additional chiral symmetry. A remarkable consequence of the correspondence is superuniversality, i.e., the ATs in some different symmetry classes of non-Hermitian systems are characterized by the same critical exponent. In addition to the comparisons between the known critical exponents for non-Hermitian systems and their Hermitian counterparts, we obtain the critical exponents in symmetry classes AI, AII, AII$^{\dagger}$, CII$^{\dagger}$, and DIII in two and three dimensions. Estimated critical exponents are consistent with the proposed correspondence. According to the correspondence, some of the exponents also give useful information of the unknown critical exponents in Hermitian systems, paving a way to study the ATs of Hermitian systems by the corresponding non-Hermitian systems.

quant-ph↗

arXiv2022

Hyperuniform electron distributions controlled by electron interactions in quasicrystal

We study how the electron-electron interactions influence the charge distributions in the metallic state of quasicrystals. As a simple theoretical model, we introduce an extended Hubbard model on the Penrose lattice, and numerically solve the model (up to $\sim1.4$ million sites) within the Hartree-Fock approximation. Because each site on the quasiperiodic lattice has a different local geometry, the Coulomb interaction, in particular the intersite one, works in a site-dependent way, leading to a nontrivial redistribution of the charge. The resultant charge distribution patterns are not multifractal but characterized by hyperuniformity, which offers a measure to distinguish various inhomogeneous but ordered distributions. We clarify how the electron interactions alter the order metric of the hyperuniformity, revealing that the intersite interaction considerably affects the hyperuniformity in particular on the electron-rich side.

cond-mat.str-el↗