Searcharxiv⌕ Search

arXiv subjects

Yi Zhang

Publications and source records attributed to Yi Zhang.

At least 793 records · Page 44Linked to original sources
arXiv2023

A Unified Framework for Solving a General Class of Nonconvexly Regularized Convex Models

Recently, several nonconvex sparse regularizers which can preserve the convexity of the cost function have received increasing attention. This paper proposes a general class of such convexity-preserving (CP) regularizers, termed partially smoothed difference-of-convex (pSDC) regularizer. The pSDC regularizer is formulated as a structured difference-of-convex (DC) function, where the landscape of the subtrahend function can be adjusted by a parameterized smoothing function so as to attain overall-convexity. Assigned with proper building blocks, the pSDC regularizer reproduces existing CP regularizers and opens the way to a large number of promising new ones. With respect to the resultant nonconvexly regularized convex (NRC) model, we derive a series of overall-convexity conditions which naturally embrace the conditions in previous works. Moreover, we develop a unified framework based on DC programming for solving the NRC model. Compared to previously reported proximal splitting type approaches, the proposed framework makes less stringent assumptions. We establish the convergence of the proposed framework to a global minimizer. Numerical experiments demonstrate the power of the pSDC regularizers and the efficiency of the proposed DC algorithm.

eess.SP↗
arXiv2023

Analysis of a mixed finite element method for stochastic Cahn-Hilliard equation with multiplicative noise

This paper proposes and analyzes a novel fully discrete finite element scheme with the interpolation operator for stochastic Cahn-Hilliard equations with functional-type noise. The nonlinear term satisfies a one-side Lipschitz condition and the diffusion term is globally Lipschitz continuous. The novelties of this paper are threefold. First, the $L^2$-stability ($L^\infty$ in time) and the discrete $H^2$-stability ($L^2$ in time) are proved for the proposed scheme. The idea is to utilize the special structure of the matrix assembled by the nonlinear term. None of these stability results has been proved for the fully implicit scheme in existing literature due to the difficulty arising from the interaction of the nonlinearity and the multiplicative noise. Second, the higher moment stability in $L^2$-norm of the discrete solution is established based on the previous stability results. Third, the Hölder continuity in time for the strong solution is established under the minimum assumption of the strong solution. Based on these, the discrete $H^{-1}$-norm of the strong convergence is discussed. Several numerical experiments including stability and convergence are also presented to validate our theoretical results.

math.NA↗
arXiv2023

Maximum-Likelihood-Estimate Hamiltonian learning via efficient and robust quantum likelihood gradient

Given the recent developments in quantum techniques, modeling the physical Hamiltonian of a target quantum many-body system is becoming an increasingly practical and vital research direction. Here, we propose an efficient strategy combining maximum likelihood estimation, gradient descent, and quantum many-body algorithms. Given the measurement outcomes, we optimize the target model Hamiltonian and density operator via a series of descents along the quantum likelihood gradient, which we prove is negative semi-definite with respect to the negative-log-likelihood function. In addition to such optimization efficiency, our maximum-likelihood-estimate Hamiltonian learning respects the locality of a given quantum system, therefore, extends readily to larger systems with available quantum many-body algorithms. Compared with previous approaches, it also exhibits better accuracy and overall stability toward noises, fluctuations, and temperature ranges, which we demonstrate with various examples.

quant-ph↗
arXiv2023

Quantum Dropout: On and Over the Hardness of Quantum Approximate Optimization Algorithm

A combinatorial optimization problem becomes very difficult in situations where the energy landscape is rugged, and the global minimum locates in a narrow region of the configuration space. When using the quantum approximate optimization algorithm (QAOA) to tackle these harder cases, we find that difficulty mainly originates from the QAOA quantum circuit instead of the cost function. To alleviate the issue, we selectively dropout the clauses defining the quantum circuit while keeping the cost function intact. Due to the combinatorial nature of the optimization problems, the dropout of clauses in the circuit does not affect the solution. Our numerical results confirm improvements in QAOA's performance with various types of quantum-dropout implementation.

quant-ph↗
arXiv2023

An XMM-Newton View of the ANdromeda Galaxy as Explored in a Legacy Survey (New-ANGELS) I: the X-ray Source Catalogue

We introduce the New-ANGELS program, an XMM-Newton survey of $\sim7.2\rm~deg^2$ area around M 31, which aims to study the X-ray populations in M 31 disk and the X-ray emitting hot gas in the inner halo of M 31 up to 30 kpc. In this first paper, we report the catalogue of 4506 detected X-ray sources, and attempt to cross-identify or roughly classify them. We identify 352 single stars in the foreground, 35 globular clusters and 27 supernova remnants associated with M 31, as well as 62 AGNs, 59 galaxies, and 1 galaxy clusters in the background. We uniquely classify 236 foreground stars and 17 supersoft sources based on their X-ray colors. X-ray binaries (83 LMXBs, 1 HMXBs) are classified based on their X-ray colors and X-ray variabilities. The remaining X-ray sources either have too low S/N to calculate their X-ray colors or do not have a unique classification, so are regarded as unclassified. The X-ray source catalogue is published online. Study of the X-ray source populations and the contribution of X-ray sources in the unresolved X-ray emissions based on this catalogue will be published in companion papers.

astro-ph.HE↗
arXiv2023

Kramers Fulde-Ferrell state and superconducting spin diode effect

We study a novel Fulde-Ferrell equal-spin pairing state with opposite center of mass Cooper pair momentum for each spin polarization. This state respects time-reversal symmetry and is dubbed a Kramers Fulde-Ferrell (KFF) state. It can be realized in a one-dimensional system with spin-orbit coupling and nearest neighbor attraction. We find that the KFF state supports nonreciprocal spin transport for both bulk superconductor and Josephson junctions when inversion symmetry is broken. In addition to the spin Josephson diode effect, the charge transport is controlled by intriguing dynamics of bound states whose transitions can be manipulated by the length of the KFF superconductor. We discuss the relevance of the effective theory to the novel physics observed in monolayer Fe-based superconductor along line defects and domain walls, and the potential for using spins to make dissipationless superconducting spintronics.

cond-mat.str-el↗
arXiv2023

(Re)$^2$H2O: Autonomous Driving Scenario Generation via Reversely Regularized Hybrid Offline-and-Online Reinforcement Learning

Autonomous driving and its widespread adoption have long held tremendous promise. Nevertheless, without a trustworthy and thorough testing procedure, not only does the industry struggle to mass-produce autonomous vehicles (AV), but neither the general public nor policymakers are convinced to accept the innovations. Generating safety-critical scenarios that present significant challenges to AV is an essential first step in testing. Real-world datasets include naturalistic but overly safe driving behaviors, whereas simulation would allow for unrestricted exploration of diverse and aggressive traffic scenarios. Conversely, higher-dimensional searching space in simulation disables efficient scenario generation without real-world data distribution as implicit constraints. In order to marry the benefits of both, it seems appealing to learn to generate scenarios from both offline real-world and online simulation data simultaneously. Therefore, we tailor a Reversely Regularized Hybrid Offline-and-Online ((Re)$^2$H2O) Reinforcement Learning recipe to additionally penalize Q-values on real-world data and reward Q-values on simulated data, which ensures the generated scenarios are both varied and adversarial. Through extensive experiments, our solution proves to produce more risky scenarios than competitive baselines and it can generalize to work with various autonomous driving models. In addition, these generated scenarios are also corroborated to be capable of fine-tuning AV performance.

cs.LG↗
arXiv2023

Belle II observation prospects for axion-like particle production from $B$ meson annihilation decay

We investigate a new production mechanism of axion-like particle (ALP) from $B$ meson annihilation decays and its observation potential at the Belle and Belle II experiments. This mechanism allows for the production of ALP from $B$ meson decays in association with a large variety of mesons. In this article, we first estimate the branching ratios of such processes with a perturbative QCD method. Focussing on the most promising $B \to h a^{\prime}$ ($h=K^\pm, π^\pm, D^0$ and $D_s$) channels, we perform sensitivity studies for $a^{\prime}$ decaying invisibly or into diphoton with Belle and Belle II experiments.

hep-ph↗
arXiv2023

Aerial Vision-and-Dialog Navigation

The ability to converse with humans and follow natural language commands is crucial for intelligent unmanned aerial vehicles (a.k.a. drones). It can relieve people's burden of holding a controller all the time, allow multitasking, and make drone control more accessible for people with disabilities or with their hands occupied. To this end, we introduce Aerial Vision-and-Dialog Navigation (AVDN), to navigate a drone via natural language conversation. We build a drone simulator with a continuous photorealistic environment and collect a new AVDN dataset of over 3k recorded navigation trajectories with asynchronous human-human dialogs between commanders and followers. The commander provides initial navigation instruction and further guidance by request, while the follower navigates the drone in the simulator and asks questions when needed. During data collection, followers' attention on the drone's visual observation is also recorded. Based on the AVDN dataset, we study the tasks of aerial navigation from (full) dialog history and propose an effective Human Attention Aided Transformer model (HAA-Transformer), which learns to predict both navigation waypoints and human attention.

cs.CV↗
arXiv2023

A Unified Conditional Framework for Diffusion-based Image Restoration

Diffusion Probabilistic Models (DPMs) have recently shown remarkable performance in image generation tasks, which are capable of generating highly realistic images. When adopting DPMs for image restoration tasks, the crucial aspect lies in how to integrate the conditional information to guide the DPMs to generate accurate and natural output, which has been largely overlooked in existing works. In this paper, we present a unified conditional framework based on diffusion models for image restoration. We leverage a lightweight UNet to predict initial guidance and the diffusion model to learn the residual of the guidance. By carefully designing the basic module and integration module for the diffusion model block, we integrate the guidance and other auxiliary conditional information into every block of the diffusion model to achieve spatially-adaptive generation conditioning. To handle high-resolution images, we propose a simple yet effective inter-step patch-splitting strategy to produce arbitrary-resolution images without grid artifacts. We evaluate our conditional framework on three challenging tasks: extreme low-light denoising, deblurring, and JPEG restoration, demonstrating its significant improvements in perceptual quality and the generalization to restoration tasks.

cs.CV↗
arXiv2023

Convergent expansions and bounds for the incomplete elliptic integral of the second kind near the logarithmic singularity

We find two series expansions for Legendre's second incomplete elliptic integral $E(λ, k)$ in terms of recursively computed elementary functions. Both expansions converge at every point of the unit square in the $(λ, k)$ plane. Partial sums of the proposed expansions form a sequence of approximations to $E(λ,k)$ which are asymptotic when $λ$ and/or $k$ tend to unity, including when both approach the logarithmic singularity $λ=k=1$ from any direction. Explicit two-sided error bounds are given at each approximation order. These bounds yield a sequence of increasingly precise asymptotically correct two-sided inequalities for $E(λ, k)$. For the reader's convenience we further present explicit expressions for low-order approximations and numerical examples to illustrate their accuracy. Our derivations are based on series rearrangements, hypergeometric summation algorithms and extensive use of the properties of the generalized hypergeometric functions including some recent inequalities.

math.CA↗
arXiv2023

Controllable Text-to-Image Generation with GPT-4

Current text-to-image generation models often struggle to follow textual instructions, especially the ones requiring spatial reasoning. On the other hand, Large Language Models (LLMs), such as GPT-4, have shown remarkable precision in generating code snippets for sketching out text inputs graphically, e.g., via TikZ. In this work, we introduce Control-GPT to guide the diffusion-based text-to-image pipelines with programmatic sketches generated by GPT-4, enhancing their abilities for instruction following. Control-GPT works by querying GPT-4 to write TikZ code, and the generated sketches are used as references alongside the text instructions for diffusion models (e.g., ControlNet) to generate photo-realistic images. One major challenge to training our pipeline is the lack of a dataset containing aligned text, images, and sketches. We address the issue by converting instance masks in existing datasets into polygons to mimic the sketches used at test time. As a result, Control-GPT greatly boosts the controllability of image generation. It establishes a new state-of-art on the spatial arrangement and object positioning generation and enhances users' control of object positions, sizes, etc., nearly doubling the accuracy of prior models. Our work, as a first attempt, shows the potential for employing LLMs to enhance the performance in computer vision tasks.

cs.CV↗
arXiv2023

Accelerating Chemical Exchange Saturation Transfer Imaging Using a Model-based Deep Neural Network With Synthetic Training Data

Purpose: To develop a model-based deep neural network for high-quality image reconstruction of undersampled multi-coil chemical exchange saturation transfer (CEST) data. Theory and Methods: Inspired by the variational network, the CEST image reconstruction equation is unrolled into a deep neural network (CEST-VN) with a k-space data-sharing block that takes advantage of the inherent redundancy in adjacent CEST frames and 3D spatial-frequential convolution kernels that exploit correlations in the x-ω domain. Additionally, a new pipeline based on multiple-pool Bloch-McConnell simulations is devised to synthesize multi-coil CEST data from publicly available anatomical MRI data. The proposed neural network is trained on simulated data with a CEST-specific loss function that jointly measures the structural and CEST contrast. The performance of CEST-VN was evaluated on three healthy volunteers and five brain tumor patients using retrospectively undersampled data with various acceleration factors, and compared with other state-of-the-art reconstruction methods. Results: The proposed CEST-VN method generated high-quality CEST source images and APT-weighted (APTw) maps in healthy and brain tumor subjects, consistently outperforming GRAPPA, blind compressed sensing, and the original variational network. With the acceleration factors increasing from 3 to 6, CEST-VN with the same hyperparameters yielded similar and accurate reconstruction without apparent loss of details or increase of artifacts. The ablation studies confirmed the effectiveness of the joint CEST-specific loss function and data-sharing block used. Conclusions: The proposed CEST-VN method can offer high-quality CEST source images and APTw maps from highly undersampled multi-coil data by integrating the deep-learning prior and multi-coil sensitivity encoding model.

physics.med-ph↗
arXiv2023

Measuring and Mitigating Constraint Violations of In-Context Learning for Utterance-to-API Semantic Parsing

In executable task-oriented semantic parsing, the system aims to translate users' utterances in natural language to machine-interpretable programs (API calls) that can be executed according to pre-defined API specifications. With the popularity of Large Language Models (LLMs), in-context learning offers a strong baseline for such scenarios, especially in data-limited regimes. However, LLMs are known to hallucinate and therefore pose a formidable challenge in constraining generated content. Thus, it remains uncertain if LLMs can effectively perform task-oriented utterance-to-API generation where respecting API's structural and task-specific constraints is crucial. In this work, we seek to measure, analyze and mitigate such constraints violations. First, we identify the categories of various constraints in obtaining API-semantics from task-oriented utterances, and define fine-grained metrics that complement traditional ones. Second, we leverage these metrics to conduct a detailed error analysis of constraints violations seen in state-of-the-art LLMs, which motivates us to investigate two mitigation strategies: Semantic-Retrieval of Demonstrations (SRD) and API-aware Constrained Decoding (API-CD). Our experiments show that these strategies are effective at reducing constraints violations and improving the quality of the generated API calls, but require careful consideration given their implementation complexity and latency.

cs.AI↗
arXiv2023

A Simple Baseline for Video Restoration with Grouped Spatial-temporal Shift

Video restoration, which aims to restore clear frames from degraded videos, has numerous important applications. The key to video restoration depends on utilizing inter-frame information. However, existing deep learning methods often rely on complicated network architectures, such as optical flow estimation, deformable convolution, and cross-frame self-attention layers, resulting in high computational costs. In this study, we propose a simple yet effective framework for video restoration. Our approach is based on grouped spatial-temporal shift, which is a lightweight and straightforward technique that can implicitly capture inter-frame correspondences for multi-frame aggregation. By introducing grouped spatial shift, we attain expansive effective receptive fields. Combined with basic 2D convolution, this simple framework can effectively aggregate inter-frame information. Extensive experiments demonstrate that our framework outperforms the previous state-of-the-art method, while using less than a quarter of its computational cost, on both video deblurring and video denoising tasks. These results indicate the potential for our approach to significantly reduce computational overhead while maintaining high-quality results. Code is avaliable at https://github.com/dasongli1/Shift-Net.

eess.IV↗
arXiv2023

Superior ferroelectricity and nonlinear optical response in a hybrid germanium iodide hexagonal perovskite

Abundant chemical diversity and structural tunability make organic-inorganic hybrid perovskites (OIHPs) a rich ore for ferroelectrics. However, compared with their inorganic counterparts such as BaTiO$_3$, their ferroelectric key properties, including large spontaneous polarization ($P_s$), low coercive field ($E_c$), and strong second harmonic generation (SHG) response, have long been great challenges, which hinder their commercial applications. Here, a quasi-one-dimensional OIHP DMAGeI$_3$ (DMA=Dimethylamine) is reported, with notable ferroelectric attributes at room temperature: a large $P_s$ of 24.14 $μ$C/cm$^2$ (on a par with BaTiO$_3$), a low $E_c$ below 2.2 kV/cm, and the strongest SHG intensity in OIHP family (about 12 times of KH$_2$PO$_4$ (KDP)). Revealed by the first-principles calculations, its large $P_s$ originates from the synergistic effects of the stereochemically active $4s^2$ lone pair of Ge$^{2+}$ and the ordering of organic cations, and its low kinetic energy barrier of small DMA cations results in a low $E_c$. Our work brings the comprehensive ferroelectric performances of OIHPs to a comparable level with commercial inorganic ferroelectric perovskites.

cond-mat.mtrl-sci↗
arXiv2023

On the structure of optimal solutions in a mathematical programming problem in a convex space

We consider an optimization problem in a convex space $E$ with an affine objective function, subject to $J$ constraints in the forms of inequalities on some other affine functions, where $J$ is a given nonnegative integer. Under suitable conditions, we apply the Feinberg-Shwartz lemma in finite dimensional convex analysis to show that there exists an optimal solution, which is in the form of a mixture of no more than $J+1$ extreme points of $E$. It seems that in the current setup, this result has not yet been made available, because the concerned problem does not fit into the framework of standard convex optimization problems.

math.OC↗
arXiv2023

From Association to Generation: Text-only Captioning by Unsupervised Cross-modal Mapping

With the development of Vision-Language Pre-training Models (VLPMs) represented by CLIP and ALIGN, significant breakthroughs have been achieved for association-based visual tasks such as image classification and image-text retrieval by the zero-shot capability of CLIP without fine-tuning. However, CLIP is hard to apply to generation-based tasks. This is due to the lack of decoder architecture and pre-training tasks for generation. Although previous works have created generation capacity for CLIP through additional language models, a modality gap between the CLIP representations of different modalities and the inability of CLIP to model the offset of this gap, which fails the concept to transfer across modalities. To solve the problem, we try to map images/videos to the language modality and generate captions from the language modality. In this paper, we propose the K-nearest-neighbor Cross-modality Mapping (Knight), a zero-shot method from association to generation. With text-only unsupervised training, Knight achieves State-of-the-Art performance in zero-shot methods for image captioning and video captioning. Our code is available at https://github.com/junyangwang0410/Knight.

cs.CV↗