Searcharxiv⌕ Search

arXiv subjects

Yuxuan Zhang

Publications and source records attributed to Yuxuan Zhang.

At least 181 records · Page 10Linked to original sources
arXiv2023

An Edge Alignment-based Orientation Selection Method for Neutron Tomography

Neutron computed tomography (nCT) is a 3D characterization technique used to image the internal morphology or chemical composition of samples in biology and materials sciences. A typical workflow involves placing the sample in the path of a neutron beam, acquiring projection data at a predefined set of orientations, and processing the resulting data using an analytic reconstruction algorithm. Typical nCT scans require hours to days to complete and are then processed using conventional filtered back-projection (FBP), which performs poorly with sparse views or noisy data. Hence, the main methods in order to reduce overall acquisition time are the use of an improved sampling strategy combined with the use of advanced reconstruction methods such as model-based iterative reconstruction (MBIR). In this paper, we propose an adaptive orientation selection method in which an MBIR reconstruction on previously-acquired measurements is used to define an objective function on orientations that balances a data-fitting term promoting edge alignment and a regularization term promoting orientation diversity. Using simulated and experimental data, we demonstrate that our method produces high-quality reconstructions using significantly fewer total measurements than the conventional approach.

eess.IV↗
arXiv2023

Video4MRI: An Empirical Study on Brain Magnetic Resonance Image Analytics with CNN-based Video Classification Frameworks

To address the problem of medical image recognition, computer vision techniques like convolutional neural networks (CNN) are frequently used. Recently, 3D CNN-based models dominate the field of magnetic resonance image (MRI) analytics. Due to the high similarity between MRI data and videos, we conduct extensive empirical studies on video recognition techniques for MRI classification to answer the questions: (1) can we directly use video recognition models for MRI classification, (2) which model is more appropriate for MRI, (3) are the common tricks like data augmentation in video recognition still useful for MRI classification? Our work suggests that advanced video techniques benefit MRI classification. In this paper, four datasets of Alzheimer's and Parkinson's disease recognition are utilized in experiments, together with three alternative video recognition models and data augmentation techniques that are frequently applied to video tasks. In terms of efficiency, the results reveal that the video framework performs better than 3D-CNN models by 5% - 11% with 50% - 66% less trainable parameters. This report pushes forward the potential fusion of 3D medical imaging and video understanding research.

cs.CV↗
arXiv2022

Fractional disclination charge and discrete shift in the Hofstadter butterfly

In the presence of crystalline symmetries, topological phases of matter acquire a host of invariants leading to non-trivial quantized responses. Here we study a particular invariant, the discrete shift $\mathscr{S}$, for the square lattice Hofstadter model of free fermions. $\mathscr{S}$ is associated with a $\mathbb{Z}_M$ classification in the presence of $M$-fold rotational symmetry and charge conservation. $\mathscr{S}$ gives quantized contributions to (i) the fractional charge bound to a lattice disclination, and (ii) the angular momentum of the ground state with an additional, symmetrically inserted magnetic flux. $\mathscr{S}$ forms its own `Hofstadter butterfly', which we numerically compute, refining the usual phase diagram of the Hofstadter model. We propose an empirical formula for $\mathscr{S}$ in terms of density and flux per plaquette for the Hofstadter bands, and we derive a number of general constraints. We show that bands with the same Chern number may have different values of $\mathscr{S}$, although odd and even Chern number bands always have half-integer and integer values of $\mathscr{S}$ respectively.

cond-mat.str-el↗
arXiv2022

An Attention-based Multi-Scale Feature Learning Network for Multimodal Medical Image Fusion

Medical images play an important role in clinical applications. Multimodal medical images could provide rich information about patients for physicians to diagnose. The image fusion technique is able to synthesize complementary information from multimodal images into a single image. This technique will prevent radiologists switch back and forth between different images and save lots of time in the diagnostic process. In this paper, we introduce a novel Dilated Residual Attention Network for the medical image fusion task. Our network is capable to extract multi-scale deep semantic features. Furthermore, we propose a novel fixed fusion strategy termed Softmax-based weighted strategy based on the Softmax weights and matrix nuclear norm. Extensive experiments show our proposed network and fusion strategy exceed the state-of-the-art performance compared with reference image fusion methods on four commonly used fusion metrics.

cs.CV↗
arXiv2022

All-photonic one-way quantum repeaters

Quantum repeater is the key technology enabler for long-distance quantum communication. To date, most of the existing quantum repeater protocols are designed based on specific quantum codes or graph states. In this paper, we propose a general framework for all-photonic one-way quantum repeaters based on the measurement-based error correction, which can be adapted to any Calderbank-Shor-Steane codes including the recently discovered quantum low density parity check (QLDPC) codes. We present a novel decoding scheme, where the error correction process is carried out at the destination based on the accumulated data from the measurements made across the network. This procedure not only outperforms the conventional protocols with independent repeaters but also simplifies the local quantum operations at repeaters. As an example, we numerically show that the [[48,6,8]] generalized bicycle code (as a small but efficient QLDPC code) has an equally good performance while reducing the resources by at least an order of magnitude.

quant-ph↗
arXiv2022

Qubit-efficient simulation of thermal states with quantum tensor networks

We present a holographic quantum simulation algorithm to variationally prepare thermal states of $d$-dimensional interacting quantum many-body systems, using only enough hardware qubits to represent a ($d$-1)-dimensional cross-section. This technique implements the thermal state by approximately unraveling the quantum matrix-product density operator (qMPDO) into a stochastic mixture of quantum matrix product states (sto-qMPS). The parameters of the quantum circuits generating the qMPS and of the probability distribution generating the stochastic mixture are determined through a variational optimization procedure. We demonstrate a small-scale proof of principle demonstration of this technique on Quantinuum's trapped-ion quantum processor to simulate thermal properties of correlated spin-chains over a wide temperature range using only a single pair of hardware qubits. Then, through classical simulations, we explore the representational power of two versions of sto-qMPS ansatzes for larger and deeper circuits and establish empirical relationships between the circuit resources and the accuracy of the variational free-energy.

quant-ph↗
arXiv2022

Holographic simulation of correlated electrons on a trapped ion quantum processor

We develop holographic quantum simulation techniques to prepare correlated electronic ground states in quantum matrix product state (qMPS) form, using far fewer qubits than the number of orbitals represented. Our approach starts with a holographic technique to prepare a compressed approximation to electronic mean-field ground-states, known as fermionic Gaussian matrix product states (GMPS), with a polynomial reduction in qubit- and (in select cases gate-) resources compared to existing techniques. Correlations are then introduced by augmenting the GMPS circuits in a variational technique which we denote GMPS+X. We demonstrate this approach on Quantinuum's System Model H1 trapped-ion quantum processor for 1$d$ models of correlated metal and Mott insulating states. Focusing on the $1d$ Fermi-Hubbard chain as a benchmark, we show that GMPS+X methods faithfully capture the physics of correlated electron states, including Mott insulators and correlated Luttinger liquid metals, using considerably fewer parameters than problem-agnostic variational circuits.

cond-mat.str-el↗
arXiv2022

The E-Bayesian Estimation and its E-MSE of Lomax distribution under different loss functions

This paper studies the E-Bayesian (expectation of the Bayesian estimation) estimation of the parameter of Lomax distribution based on different loss functions. Under different loss functions, we calculate the Bayesian estimation of the parameter and then calculate the expectation of the estimated value to get the E-Bayesian estimation. To measure the estimated error, the E-MSE (expected mean squared error) is introduced. And the formulas of E-Bayesian estimation and E-MSE are given. By applying Markov Chain Monte Carlo technology, we analyze the performances of the proposed methods. Results are compared on the basis of E-MSE. Then, cases of samples in real data sets are presented for illustration. In order to test whether the Lomax distribution can be used in analyzing the datasets, Kolmogorov Smirnov tests are conducted. Using real data, we can get the maximum likelihood estimation at the same time and compare it with E-Bayesian estimation. At last, we get the results of the comparison between Bayesian and E-Bayesian estimation methods under three different loss functions.

stat.ME↗
arXiv2022

Straddling-gates problem in multipartite quantum systems

We study a variant of quantum circuit complexity, the binding complexity: Consider a $n$-qubit system divided into two sets of $k_1$, $k_2$ qubits each ($k_1\leq k_2$) and gates within each set are free; what is the least cost of two-qubit gates ''straddling'' the sets for preparing an arbitrary quantum state, assuming no ancilla qubits allowed? Firstly, our work suggests that, without making assumptions on the entanglement spectrum, $Θ(2^{k_1})$ straddling gates always suffice. We then prove any $\text{U}(2^n)$ unitary synthesis can be accomplished with $Θ(4^{k_1})$ straddling gates. Furthermore, we extend our results to multipartite systems, and show that any $m$-partite Schmidt decomposable state has binding complexity linear in $m$, which hints its multi-separable property. This result not only resolves an open problem posed by Vijay Balasubramanian, who was initially motivated by the ''Complexity=Volume'' conjecture in quantum gravity, but also offers realistic applications in distributed quantum computation in the near future.

quant-ph↗
arXiv2022

The clamped intensity of femtosecond laser pulses varying with gas pressure in the presence of external focusing

We perform a theoretical investigation of the clamped laser intensity inside the filament plasma as a function of gas pressure with external focusing. Unlike the clamped intensity under the selffocusing condition, which is independent on the gas pressure, the clamped intensity with external focusing decreases with the gas pressure. Our findings can explain the changes of the signals of femtosecond-laser-induced 391-nm forward emission and fluorescence with the nitrogen gas pressure.

physics.plasm-ph↗
arXiv2022

A Scintillator Attenuation Spectrometer For Intense Gamma-Rays

A new type of compact high resolution high sensitivity gamma ray spectrometer for short pulse intense 250 keV to 50 MeV gamma rays has been developed by combining the principles of scintillators and attenuation spectrometers. The first prototype of this scintillator attenuation spectrometer or SAS was tested successfully in Trident laser experiments at LANL. Later versions have been used extensively in the Texas Petawatt laser experiments in Austin TX, and more recently in OMEGAEP laser experiments at LLE, Rochester, NY. The SAS is particularly useful for high repetition rate laser applications. Here we give a concise description of the design principles, capabilities and sample preliminary results of the SAS.

physics.ins-det↗
arXiv2022

The Implicit Values of A Good Hand Shake: Handheld Multi-Frame Neural Depth Refinement

Modern smartphones can continuously stream multi-megapixel RGB images at 60Hz, synchronized with high-quality 3D pose information and low-resolution LiDAR-driven depth estimates. During a snapshot photograph, the natural unsteadiness of the photographer's hands offers millimeter-scale variation in camera pose, which we can capture along with RGB and depth in a circular buffer. In this work we explore how, from a bundle of these measurements acquired during viewfinding, we can combine dense micro-baseline parallax cues with kilopixel LiDAR depth to distill a high-fidelity depth map. We take a test-time optimization approach and train a coordinate MLP to output photometrically and geometrically consistent depth estimates at the continuous coordinates along the path traced by the photographer's natural hand shake. With no additional hardware, artificial hand motion, or user interaction beyond the press of a button, our proposed method brings high-resolution depth estimates to point-and-shoot "tabletop" photography -- textured objects at close range.

cs.CV↗
arXiv2022

All You Need is RAW: Defending Against Adversarial Attacks with Camera Image Pipelines

Existing neural networks for computer vision tasks are vulnerable to adversarial attacks: adding imperceptible perturbations to the input images can fool these methods to make a false prediction on an image that was correctly predicted without the perturbation. Various defense methods have proposed image-to-image mapping methods, either including these perturbations in the training process or removing them in a preprocessing denoising step. In doing so, existing methods often ignore that the natural RGB images in today's datasets are not captured but, in fact, recovered from RAW color filter array captures that are subject to various degradations in the capture. In this work, we exploit this RAW data distribution as an empirical prior for adversarial defense. Specifically, we proposed a model-agnostic adversarial defensive method, which maps the input RGB images to Bayer RAW space and back to output RGB using a learned camera image signal processing (ISP) pipeline to eliminate potential adversarial patterns. The proposed method acts as an off-the-shelf preprocessing module and, unlike model-specific adversarial training methods, does not require adversarial images to train. As a result, the method generalizes to unseen tasks without additional retraining. Experiments on large-scale datasets (e.g., ImageNet, COCO) for different vision tasks (e.g., classification, semantic segmentation, object detection) validate that the method significantly outperforms existing methods across task domains.

cs.CV↗
arXiv2022

Holographic quantum simulation of entanglement renormalization circuits

While standard approaches to quantum simulation require a number of qubits proportional to the number of simulated particles, current noisy quantum computers are limited to tens of qubits. With the technique of holographic quantum simulation, a $D$-dimensional system can be simulated with a $D{\rm -}1$-dimensional subset of qubits, enabling the study of systems significantly larger than current quantum computers. Using circuits derived from the multiscale entanglement renormalization ansatz (MERA), we accurately prepare the ground state of an $L=32$ critical, non-integrable perturbed Ising model and measure long-range correlations on the 10 qubit Quantinuum trapped ion computer. We introduce generalized MERA (gMERA) networks that interpolate between MERA and matrix product state networks and demonstrate that gMERA can capture far longer correlations than a MERA with the same number of qubits, at the expense of greater circuit depth. Finally, we perform noisy simulations of these two network ansätze and find that the optimal choice of network depends on noise level, available qubits, and the state to be represented.

quant-ph↗
arXiv2021

Graph-based multiple change-point detection

We propose a new multiple change-point detection framework for multivariate and non-Euclidean data. First, we combine graph-based statistics with wild binary segmentation or seeded binary segmentation to search for a pool of candidate change-points. We then prune the candidate change-points through a novel goodness-of-fit statistic. Numerical studies show that this new framework outperforms existing methods under a wide range of settings. The resulting change-points can further be arranged hierarchically based on the goodness-of-fit statistic. The new framework is illustrated on a Neuropixels recording of an awake mouse.

stat.ME↗
arXiv2021

QED driven QAOA for network-flow optimization

We present a general framework for modifying quantum approximate optimization algorithms (QAOA) to solve constrained network flow problems. By exploiting an analogy between flow constraints and Gauss's law for electromagnetism, we design lattice quantum electrodynamics (QED) inspired mixing Hamiltonians that preserve flow constraints throughout the QAOA process. This results in an exponential reduction in the size of the configuration space that needs to be explored, which we show through numerical simulations, yields higher quality approximate solutions compared to the original QAOA routine. We outline a specific implementation for edge-disjoint path (EDP) problems related to traffic congestion minimization, numerically analyze the effect of initial state choice, and explore trade-offs between circuit complexity and qubit resources via a particle-vortex duality mapping. Comparing the effect of initial states reveals that starting with an ergodic (unbiased) superposition of solutions yields better performance than beginning with the mixer ground-state, suggesting a departure from the "short-cut to adiabaticity" mechanism often used to motivate QAOA.

quant-ph↗
arXiv2021

Physics-driven Learning of the Steady Navier-Stokes Equations using Deep Convolutional Neural Networks

Recently, physics-driven deep learning methods have shown particular promise for the prediction of physical fields, especially to reduce the dependency on large amounts of pre-computed training data. In this work, we target the physics-driven learning of complex flow fields with high resolutions. We propose the use of \emph{Convolutional neural networks} (CNN) based U-net architectures to efficiently represent and reconstruct the input and output fields, respectively. By introducing Navier-Stokes equations and boundary conditions into loss functions, the physics-driven CNN is designed to predict corresponding steady flow fields directly. In particular, this prevents many of the difficulties associated with approaches employing fully connected neural networks. Several numerical experiments are conducted to investigate the behavior of the CNN approach, and the results indicate that a first-order accuracy has been achieved. Specifically for the case of a flow around a cylinder, different flow regimes can be learned and the adhered "twin-vortices" are predicted correctly. The numerical results also show that the training for multiple cases is accelerated significantly, especially for the difficult cases at low Reynolds numbers, and when limited reference solutions are used as supplementary learning targets.

physics.flu-dyn↗