arXiv subjects

Yue Zhao

Publications and source records attributed to Yue Zhao.

At least 289 records · Page 16Linked to original sources

arXiv2024

Stability for a multi-frequency inverse random source problem

We present stability estimates for the inverse source problem of the stochastic Helmholtz equation in two and three dimensions by either near-field or far-field data. The random source is assumed to be a microlocally isotropic generalized Gaussian random function. For the direct problem, by exploring the regularity of the Green function, we demonstrate that the direct problem admits a unique bounded solution with an explicit integral representation, which enhances the existing regularity result. For the case using near-field data, the analysis of the inverse problem employs microlocal analysis to achieve an estimate for the Fourier transform of the micro-correlation strength by the near-field correlation data and a high-frequency tail. The stability follows by showing the analyticity of the data and applying a novel analytic continuation principle. The stability estimate by far-field data is derived by investigating the correlation of the far-field data. The stability estimate consists of the Lipschitz type data discrepancy and the logarithmic stability. The latter decreases as the upper bound of the frequency increases, which exhibits the phenomenon of increasing stability.

math.AP↗

arXiv2024

Dissipative stabilization of high-dimensional GHZ states for neutral atoms

High-dimensional quantum entanglement characterizes the entanglement of quantum systems within a larger Hilbert space, introducing more intricate and complex correlations among the entangled particles' states. The high-dimensional Greenberger-Horne-Zeilinger (GHZ) state, symbolic of this type of entanglement, is of significant importance in various quantum information processing applications. This study proposes integrating a neutral atom platform with quantum reservoir engineering to generate a high-dimensional GHZ state deterministically. Leveraging the advantages of neutral atoms in a modified unconventional Rydberg pumping mechanism, combined with controlled dissipation, we achieve a three-dimensional GHZ state with a fidelity surpassing 99\% through multiple pump and dissipation cycles. This innovative approach paves the way for experimentally feasible, deterministic preparation of high-dimensional GHZ states in Rydberg atom systems, thereby advancing the capabilities of quantum information processing.

quant-ph↗

arXiv2024

Direct and Inverse scattering in a three-dimensional planar waveguide

In this paper, we study the direct and inverse scattering of the Schrödinger equation in a three-dimensional planar waveguide. For the direct problem, we derive a resonance-free region and resolvent estimates for the resolvent of the Schrödinger operator in such a geometry. Based on the analysis of the resolvent, several inverse problems are investigated. First, given the potential function, we prove the uniqueness of the inverse source problem with multi-frequency data. We also develop a Fourier-based method to reconstruct the source function. The capability of this method is numerically illustrated by examples. Second, the uniqueness and increased stability of an inverse potential problem from data generated by incident waves are achieved in the absence of the source function. To derive the stability estimate, we use an argument of quantitative analytic continuation in complex theory. Third, we prove the uniqueness of simultaneously determining the source and potential by active boundary data generated by incident waves. In these inverse problems, we only use the limited lateral Dirichlet boundary data at multiple wavenumbers within a finite interval.

math.AP↗

arXiv2024

Time-harmonic elastic scattering by unbounded deterministic and random rough surfaces in three dimensions

In this paper, we investigate well-posedness of time-harmonic scattering of elastic waves by unbounded rigid rough surfaces in three dimensions. The elastic scattering is caused by an $L^2$ function with a compact support in the $x_3$-direction, and both deterministic and random surfaces are investigated via the variational approach. The rough surface in a deterministic setting is assumed to be Lipschitz and lie within a finite distance of a flat plane, and the scattering is caused by an inhomogeneous term in the elastic wave equation whose support lies within some finite distance of the boundary. For the deterministic case, a stability estimate of elastic scattering by rough surface is shown at an arbitrary frequency. It is noticed that all constants in {\it a priori} bounds are bounded by explicit functions of the frequency and geometry of rough surfaces. Furthermore, based on this explicit dependence on the frequency together with the measurability and $\mathbb{P}$-essentially separability of the randomness, we obtain a similar bound for the solution of the scattering by random surfaces.

math.AP↗

arXiv2024

Forging Vision Foundation Models for Autonomous Driving: Challenges, Methodologies, and Opportunities

The rise of large foundation models, trained on extensive datasets, is revolutionizing the field of AI. Models such as SAM, DALL-E2, and GPT-4 showcase their adaptability by extracting intricate patterns and performing effectively across diverse tasks, thereby serving as potent building blocks for a wide range of AI applications. Autonomous driving, a vibrant front in AI applications, remains challenged by the lack of dedicated vision foundation models (VFMs). The scarcity of comprehensive training data, the need for multi-sensor integration, and the diverse task-specific architectures pose significant obstacles to the development of VFMs in this field. This paper delves into the critical challenge of forging VFMs tailored specifically for autonomous driving, while also outlining future directions. Through a systematic analysis of over 250 papers, we dissect essential techniques for VFM development, including data preparation, pre-training strategies, and downstream task adaptation. Moreover, we explore key advancements such as NeRF, diffusion models, 3D Gaussian Splatting, and world models, presenting a comprehensive roadmap for future research. To empower researchers, we have built and maintained https://github.com/zhanghm1995/Forge_VFM4AD, an open-access repository constantly updated with the latest advancements in forging VFMs for autonomous driving.

cs.CV↗

arXiv2024

Inverse source problem of the biharmonic equation from multi-frequency phaseless data

This work deals with an inverse source problem for the biharmonic wave equation. A two-stage numerical method is proposed to identify the unknown source from the multi-frequency phaseless data. In the first stage, we introduce some artificially auxiliary point sources to the inverse source system and establish a phase retrieval formula. Theoretically, we point out that the phase can be uniquely determined and estimate the stability of this phase retrieval approach. Once the phase information is retrieved, the Fourier method is adopted to reconstruct the source function from the phased multi-frequency data. The proposed method is easy-to-implement and there is no forward solver involved in the reconstruction. Numerical experiments are conducted to verify the performance of the proposed method.

math.NA↗

arXiv2024

Exploration of faint X-ray and radio sources in the massive globular cluster M14: A UV-bright counterpart to Nova Ophiuchus 1938

Using a 12 ks archival Chandra X-ray Observatory ACIS-S observation on the massive globular cluster (GC) M14, we detect a total of 7 faint X-ray sources within its half-light radius at a 0.5-7 keV depth of $2.5\times 10^{31}\,\mathrm{erg~s^{-1}}$. We cross-match the X-ray source positions with a catalogue of the Very Large Array radio point sources and a Hubble Space Telescope (HST) UV/optical/near-IR photometry catalogue, revealing radio counterparts to 2 and HST counterparts to 6 of the X-ray sources. In addition, we also identify a radio source with the recently discovered millisecond pulsar PSR 1737-0314A. The brightest X-ray source, CX1, appears to be consistent with the nominal position of the classic nova Ophiuchi 1938 (Oph 1938), and both Oph 1938 and CX1 are consistent with a UV-bright variable HST counterpart, which we argue to be the source of the nova eruption in 1938. This makes Oph 1938 the second classic nova recovered in a Galactic GC since Nova T Scorpii in M80. CX2 is consistent with the steep-spectrum radio source VLA8, which unambiguously matches a faint blue source; the steepness of VLA8 is suggestive of a pulsar nature, possibly a transitional millisecond pulsar with a late K dwarf companion, though an active galactic nucleus (AGN) cannot be ruled out. The other counterparts to the X-ray sources are all suggestive of chromospherically active binaries or background AGNs, so their nature requires further membership information.

astro-ph.HE↗

arXiv2023

Energy Stable Scheme for Random Batch Molecular Dynamics

The computational bottleneck of molecular dynamics is the pairwise additive long-range interactions between particles. The random batch Ewald (RBE) method provides a highly efficient and superscalable solver for long-range interactions, but the stochastic nature of this algorithm leads to unphysical self-heating effect during the simulation. We propose an energy stable scheme (ESS) for particle systems by employing a Berendsen-type energy bath. The scheme removes the notorious energy drift which exists due to the force error even when a symplectic integrator is employed. Combining the RBE and the ESS, the new method provides a perfect solution of the computational bottleneck of molecular dynamics at the microcanonical ensemble. Numerical results for primitive electrolyte and all-atom pure water systems demonstrate the attractive performance of the algorithm including its dramatically high accuracy, linear complexity and overcoming the energy drift for long-time simulations.

physics.chem-ph↗

arXiv2023

ADGym: Design Choices for Deep Anomaly Detection

Deep learning (DL) techniques have recently found success in anomaly detection (AD) across various fields such as finance, medical services, and cloud computing. However, most of the current research tends to view deep AD algorithms as a whole, without dissecting the contributions of individual design choices like loss functions and network architectures. This view tends to diminish the value of preliminary steps like data preprocessing, as more attention is given to newly designed loss functions, network architectures, and learning paradigms. In this paper, we aim to bridge this gap by asking two key questions: (i) Which design choices in deep AD methods are crucial for detecting anomalies? (ii) How can we automatically select the optimal design choices for a given AD dataset, instead of relying on generic, pre-existing solutions? To address these questions, we introduce ADGym, a platform specifically crafted for comprehensive evaluation and automatic selection of AD design elements in deep methods. Our extensive experiments reveal that relying solely on existing leading methods is not sufficient. In contrast, models developed using ADGym significantly surpass current state-of-the-art techniques.

cs.LG↗

arXiv2023

Superconducting Diode Effect and Large Magnetochiral Anisotropy in T$_d$-MoTe$_2$ Thin Film

In the absence of time-reversal invariance, metals without inversion symmetry may exhibit nonreciprocal charge transport -- a magnetochiral anisotropy that manifests as unequal electrical resistance for opposite current flow directions. If superconductivity also sets in, the charge transmission may become dissipationless in one direction while remaining dissipative in the opposite, thereby realizing a superconducting diode. Through both direct-current and alternating-current measurements, we study the nonreciprocal effects in thin films of the noncentrosymmetric superconductor T$_d$-MoTe\textsubscript{2} with disorders. We observe nonreciprocal superconducting critical currents with a diode efficiency close to 20\%~, and a large magnetochiral anisotropy coefficient up to $\SI{5.9e8}{\per\tesla\per\ampere}$, under weak out-of-plane magnetic field in the millitesla range. The great enhancement of rectification efficiency under out-of-plane magnetic field is likely abscribed to the vortex ratchet effect, which naturally appears in the noncentrosymmetric superconductor with disorders. Intriguingly, unlike the finding in Rashba systems, the strongest in-plane nonreciprocal effect does not occur when the field is perpendicular to the current flow direction. We develop a phenomenological theory to demonstrate that this peculiar behavior can be attributed to the asymmetric structure of spin-orbit coupling in T$_d$-MoTe\textsubscript{2}. Our study highlights how the crystallographic symmetry critically impacts the nonreciprocal transport, and would further advance the research for designing the superconducting diode with the best performance.

cond-mat.supr-con↗

arXiv2023

On the natal kick of the black hole X-ray binary H 1705--250

When a compact object is formed, an impulse (kick) will be imparted to the system by the mass lost during the core-collapse supernova (SN). A number of other mechanisms may impart an additional kick on the system, although evidence for these natal kicks in black hole systems remains limited. Updated Gaia astrometry has recently identified a number of high peculiar velocity (in excess of Galactic motion) compact objects. Here, we focus on the black hole low-mass X-ray binary H 1705--250, which has a peculiar velocity $\upsilon_{\mathrm{pec}}\,=\,221^{+101}_{-108}\,\mathrm{km}\,\mathrm{s}^{-1}$. Using population synthesis to reconstruct its evolutionary history (assuming formation via isolated binary evolution within the Galactic plane), we constrain the properties of the progenitor and pre-SN orbit. The magnitude of a kick solely due to mass loss is found to be $\sim\,30\,\mathrm{km}\,\mathrm{s}^{-1}$, which cannot account for the high present-day peculiar motion. We therefore deduce that the black hole received an additional natal kick at formation, and place limits on its magnitude, finding it to be $\sim\,295\,\mathrm{km}\,\mathrm{s}^{-1}$ (minimum $90\,\mathrm{km}\,\mathrm{s}^{-1}$). This furthers the argument that these kicks are not limited to neutron stars.

astro-ph.HE↗

arXiv2023

What Makes for Robust Multi-Modal Models in the Face of Missing Modalities?

With the growing success of multi-modal learning, research on the robustness of multi-modal models, especially when facing situations with missing modalities, is receiving increased attention. Nevertheless, previous studies in this domain exhibit certain limitations, as they often lack theoretical insights or their methodologies are tied to specific network architectures or modalities. We model the scenarios of multi-modal models encountering missing modalities from an information-theoretic perspective and illustrate that the performance ceiling in such scenarios can be approached by efficiently utilizing the information inherent in non-missing modalities. In practice, there are two key aspects: (1) The encoder should be able to extract sufficiently good features from the non-missing modality; (2) The extracted features should be robust enough not to be influenced by noise during the fusion process across modalities. To this end, we introduce Uni-Modal Ensemble with Missing Modality Adaptation (UME-MMA). UME-MMA employs uni-modal pre-trained weights for the multi-modal model to enhance feature extraction and utilizes missing modality data augmentation techniques to better adapt to situations with missing modalities. Apart from that, UME-MMA, built on a late-fusion learning framework, allows for the plug-and-play use of various encoders, making it suitable for a wide range of modalities and enabling seamless integration of large-scale pre-trained encoders to further enhance performance. And we demonstrate UME-MMA's effectiveness in audio-visual datasets~(e.g., AV-MNIST, Kinetics-Sound, AVE) and vision-language datasets~(e.g., MM-IMDB, UPMC Food101).

cs.AI↗

arXiv2023

Improving Discriminative Multi-Modal Learning with Large-Scale Pre-Trained Models

This paper investigates how to better leverage large-scale pre-trained uni-modal models to further enhance discriminative multi-modal learning. Even when fine-tuned with only uni-modal data, these models can outperform previous multi-modal models in certain tasks. It's clear that their incorporation into multi-modal learning would significantly improve performance. However, multi-modal learning with these models still suffers from insufficient learning of uni-modal features, which weakens the resulting multi-modal model's generalization ability. While fine-tuning uni-modal models separately and then aggregating their predictions is straightforward, it doesn't allow for adequate adaptation between modalities, also leading to sub-optimal results. To this end, we introduce Multi-Modal Low-Rank Adaptation learning (MMLoRA). By freezing the weights of uni-modal fine-tuned models, adding extra trainable rank decomposition matrices to them, and subsequently performing multi-modal joint training, our method enhances adaptation between modalities and boosts overall performance. We demonstrate the effectiveness of MMLoRA on three dataset categories: audio-visual (e.g., AVE, Kinetics-Sound, CREMA-D), vision-language (e.g., MM-IMDB, UPMC Food101), and RGB-Optical Flow (UCF101).

cs.CV↗

arXiv2023

LEAP: Liberate Sparse-view 3D Modeling from Camera Poses

Are camera poses necessary for multi-view 3D modeling? Existing approaches predominantly assume access to accurate camera poses. While this assumption might hold for dense views, accurately estimating camera poses for sparse views is often elusive. Our analysis reveals that noisy estimated poses lead to degraded performance for existing sparse-view 3D modeling methods. To address this issue, we present LEAP, a novel pose-free approach, therefore challenging the prevailing notion that camera poses are indispensable. LEAP discards pose-based operations and learns geometric knowledge from data. LEAP is equipped with a neural volume, which is shared across scenes and is parameterized to encode geometry and texture priors. For each incoming scene, we update the neural volume by aggregating 2D image features in a feature-similarity-driven manner. The updated neural volume is decoded into the radiance field, enabling novel view synthesis from any viewpoint. On both object-centric and scene-level datasets, we show that LEAP significantly outperforms prior methods when they employ predicted poses from state-of-the-art pose estimators. Notably, LEAP performs on par with prior approaches that use ground-truth poses while running $400\times$ faster than PixelNeRF. We show LEAP generalizes to novel object categories and scenes, and learns knowledge closely resembles epipolar geometry. Project page: https://hwjiang1510.github.io/LEAP/

cs.CV↗

arXiv2023

Reformulating Vision-Language Foundation Models and Datasets Towards Universal Multimodal Assistants

Recent Multimodal Large Language Models (MLLMs) exhibit impressive abilities to perceive images and follow open-ended instructions. The capabilities of MLLMs depend on two crucial factors: the model architecture to facilitate the feature alignment of visual modules and large language models; the multimodal instruction tuning datasets for human instruction following. (i) For the model architecture, most existing models introduce an external bridge module to connect vision encoders with language models, which needs an additional feature-alignment pre-training. In this work, we discover that compact pre-trained vision language models can inherently serve as ``out-of-the-box'' bridges between vision and language. Based on this, we propose Muffin framework, which directly employs pre-trained vision-language models to act as providers of visual signals. (ii) For the multimodal instruction tuning datasets, existing methods omit the complementary relationship between different datasets and simply mix datasets from different tasks. Instead, we propose UniMM-Chat dataset which explores the complementarities of datasets to generate 1.1M high-quality and diverse multimodal instructions. We merge information describing the same image from diverse datasets and transforms it into more knowledge-intensive conversation data. Experimental results demonstrate the effectiveness of the Muffin framework and UniMM-Chat dataset. Muffin achieves state-of-the-art performance on a wide range of vision-language tasks, significantly surpassing state-of-the-art models like LLaVA and InstructBLIP. Our model and dataset are all accessible at https://github.com/thunlp/muffin.

cs.CV↗

arXiv2023

Training a Large Video Model on a Single Machine in a Day

Videos are big, complex to pre-process, and slow to train on. State-of-the-art large-scale video models are trained on clusters of 32 or more GPUs for several days. As a consequence, academia largely ceded the training of large video models to industry. In this paper, we show how to still train a state-of-the-art video model on a single machine with eight consumer-grade GPUs in a day. We identify three bottlenecks, IO, CPU, and GPU computation, and optimize each. The result is a highly efficient video training pipeline. For comparable architectures, our pipeline achieves higher accuracies with $\frac{1}{8}$ of the computation compared to prior work. Code is available at https://github.com/zhaoyue-zephyrus/AVION.

cs.CV↗

arXiv2023

Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystack

Training text-to-image models with web scale image-text pairs enables the generation of a wide range of visual concepts from text. However, these pre-trained models often face challenges when it comes to generating highly aesthetic images. This creates the need for aesthetic alignment post pre-training. In this paper, we propose quality-tuning to effectively guide a pre-trained model to exclusively generate highly visually appealing images, while maintaining generality across visual concepts. Our key insight is that supervised fine-tuning with a set of surprisingly small but extremely visually appealing images can significantly improve the generation quality. We pre-train a latent diffusion model on $1.1$ billion image-text pairs and fine-tune it with only a few thousand carefully selected high-quality images. The resulting model, Emu, achieves a win rate of $82.9\%$ compared with its pre-trained only counterpart. Compared to the state-of-the-art SDXLv1.0, Emu is preferred $68.4\%$ and $71.3\%$ of the time on visual appeal on the standard PartiPrompts and our Open User Input benchmark based on the real-world usage of text-to-image models. In addition, we show that quality-tuning is a generic approach that is also effective for other architectures, including pixel diffusion and masked generative transformer models.

cs.CV↗

arXiv2023

Accurate 3D Prediction of Missing Teeth in Diverse Patterns for Precise Dental Implant Planning

In recent years, the demand for dental implants has surged, driven by their high success rates and esthetic advantages. However, accurate prediction of missing teeth for precise digital implant planning remains a challenge due to the intricate nature of dental structures and the variability in tooth loss patterns. This study presents a novel framework for accurate prediction of missing teeth in different patterns, facilitating digital implant planning. The proposed framework begins by estimating point-to-point correspondence among a dataset of dental mesh models reconstructed from CBCT images of healthy subjects. Subsequently, tooth dictionaries are constructed for each tooth type, encoding their position and shape information based on the established point-to-point correspondence. To predict missing teeth in a given dental mesh model, sparse coefficients are learned by sparsely representing adjacent teeth of the missing teeth using the corresponding tooth dictionaries. These coefficients are then applied to the dictionaries of the missing teeth to generate accurate predictions of their positions and shapes. The evaluation results on real subjects shows that our proposed framework achieves an average prediction error of 1.04mm for predictions of single missing tooth and an average prediction error of 1.33mm for the prediction of 14 missing teeth, which demonstrates its capability of accurately predicting missing teeth in various patterns. By accurately predicting missing teeth, dental professionals can improve the planning and placement of dental implants, leading to better esthetic and functional outcomes for patients undergoing dental implant procedures.

cs.CV↗