arXiv subjects

Li Wang

Publications and source records attributed to Li Wang.

At least 181 records · Page 10Linked to original sources

arXiv2025

SF-Speech: Straightened Flow for Zero-Shot Voice Clone

Recently, neural ordinary differential equations (ODE) models trained with flow matching have achieved impressive performance on the zero-shot voice clone task. Nevertheless, postulating standard Gaussian noise as the initial distribution of ODE gives rise to numerous intersections within the fitted targets of flow matching, which presents challenges to model training and enhances the curvature of the learned generated trajectories. These curved trajectories restrict the capacity of ODE models for generating desirable samples with a few steps. This paper proposes SF-Speech, a novel voice clone model based on ODE and in-context learning. Unlike the previous works, SF-Speech adopts a lightweight multi-stage module to generate a more deterministic initial distribution for ODE. Without introducing any additional loss function, we effectively straighten the curved reverse trajectories of the ODE model by jointly training it with the proposed module. Experiment results on datasets of various scales show that SF-Speech outperforms the state-of-the-art zero-shot TTS methods and requires only a quarter of the solver steps, resulting in a generation speed approximately 3.7 times that of Voicebox and E2 TTS. Audio samples are available at the demo page\footnote{[Online] Available: https://lixuyuan102.github.io/Demo/}.

cs.SD↗

arXiv2025

Enhanced TM-Mode 3D Coupled Wave Theory for Photonic Crystal Surface-Emitting Terahertz Quantum Cascade Lasers

In this study, we propose and develop an enhanced three-dimensional coupled wave theory (3D CWT) to investigate the optical field behavior in photonic crystal surface-emitting terahertz quantum cascade lasers (THz-QCLs). By incorporating an effective permittivity enhancement (EP) model and a self-consistent iteration (SCI) method, we successfully address the numerical dispersion issues encountered in analytical methods when dealing with metallic waveguide structures. The results demonstrate that the EP and SCI-enhanced 3D TM mode CWT achieves computational accuracy comparable to traditional numerical simulation methods such as finite-difference time-domain (FDTD), while significantly reducing the required computational resources, including time and memory, to just tens of minutes. Moreover, this method provides a clear physical insight, revealing the reasons behind the current low extraction efficiency in surface-emitting THz-QCLs. Our study showcases the potential of the EP and SCI-enhanced 3D CWT as a powerful simulation tool in the research of photonic crystal surface-emitting lasers, offering a new theoretical foundation and optimization direction for future laser designs.

physics.optics↗

arXiv2025

Unveiling Bifurcated Blue Straggler Sequences in NGC 2173: Insights from Binary Evolution

Identifying bifurcated blue straggler (BS) sequences in color-magnitude diagrams (CMDs) of star clusters has long been regarded as a powerful diagnostic for distinguishing different BS formation mechanisms. While such bifurcations are typically associated with core-collapsed clusters, their detection in dynamically young clusters raises new questions about their origins. In this study, using high-precision proper motion data derived from Hubble Space Telescope multi-epoch observations, we confirm the existence of two distinct BS sequences in the Large Magellanic Cloud cluster NGC 2173 (~ 1.58 Gyr): a well-defined, narrow blue sequence and a sparser red sequence. The extended core region excludes collisional formation as a viable channel for BS formation. Our binary evolution models suggest that non-conservative mass transfer (MT) is essential for shaping the distribution of BS binaries in the CMD. The red sequence primarily comprises BS binaries formed through conservative ongoing MT. In contrast, the blue sequence BS binaries are formed through non-conservative post-MT processes, all involving white dwarf (WD) companions. These BS+WD binary systems may subsequently undergo a second MT phase, leading to the formation of double WD systems.

astro-ph.SR↗

arXiv2025

GraphBEV: Towards Robust BEV Feature Alignment for Multi-Modal 3D Object Detection

Integrating LiDAR and camera information into Bird's-Eye-View (BEV) representation has emerged as a crucial aspect of 3D object detection in autonomous driving. However, existing methods are susceptible to the inaccurate calibration relationship between LiDAR and the camera sensor. Such inaccuracies result in errors in depth estimation for the camera branch, ultimately causing misalignment between LiDAR and camera BEV features. In this work, we propose a robust fusion framework called Graph BEV. Addressing errors caused by inaccurate point cloud projection, we introduce a Local Align module that employs neighbor-aware depth features via Graph matching. Additionally, we propose a Global Align module to rectify the misalignment between LiDAR and camera BEV features. Our Graph BEV framework achieves state-of-the-art performance, with an mAP of 70.1\%, surpassing BEV Fusion by 1.6\% on the nuscenes validation set. Importantly, our Graph BEV outperforms BEV Fusion by 8.3\% under conditions with misalignment noise.

cs.CV↗

arXiv2025

Reconstruction of heat relaxation index in phonon transport equation

For nano-materials, heat conductivity is an ill-defined concept. This classical concept assumes the validity of Fourier's law, which states the heat flux is proportional to temperature gradient, with heat conductivity used to denote this ratio. However, this macroscopic constitutive relation breaks down at nano-scales. Instead, heat is propagated using phonon transport equation, an ab initio model derived from the first principle. In this equation, a material's thermal property is coded in a coefficient termed the relaxation time ($τ$). We study an inverse problem in this paper, by using material's temperature response upon heat injection to infer the relaxation time. This inverse problem is formulated in a PDE-constrained optimization, and numerically solved by Stochastic Gradient Descent (SGD) method and its variants. In the execution of SGD, Fréchet derivative is computed and Lipschitz continuity is proved. This approach, in comparison to the earlier studies, honors the nano-structure of of heat conductivity in a nano-material, and we numerically verify the break down of the Fourier's law.

math.NA↗

arXiv2025

An NEPv Approach for Feature Selection via Orthogonal OCCA with the (2,1)-norm Regularization

A novel feature selection model via orthogonal canonical correlation analysis with the $(2,1)$-norm regularization is proposed, and the model is solved by a practical NEPv approach (nonlinear eigenvalue problem with eigenvector dependency), yielding a feature selection method named OCCA-FS. It is proved that OCCA-FS always produces a sequence of approximations with monotonic objective values and is globally convergent. Extensive numerical experiments are performed to compare OCCA-FS against existing feature selection methods. The numerical results demonstrate that OCCA-FS produces superior classification performance and often comes out on the top among all feature selection methods in comparison.

math.NA↗

arXiv2025

Multiple stellar populations at less evolved stages. IV. evidence of helium enrichments in four Magellanic globular clusters

Most globular clusters in the local group show multiple stellar populations, a pattern defined by variations of several light elements, with some also displaying dispersions in heavy elements. Since helium is the most immediate product of hydrogen burning, almost all models suggest that second-generation stars should show different levels of helium enrichment than first-generation stars. Therefore, investigating helium enrichment in stellar populations of globular clusters can constrain different theoretical models. Using the deep photometry carried out by the {\sl Hubble} Space Telescope, we have analyzed the morphologies of the main sequences of four Large Magellanic Cloud globular clusters, Hodge 11, NGC 1841, NGC 2210, and NGC 2257. We aim to constrain the helium distributions among their main sequence stars. We found a clear broadening of main sequences in all four clusters. After excluding the effects of photometric errors, differential reddening, unresolved binaries, and metallicity dispersions, this broadening would suggest a substantial helium enrichment in their star populations, ranging from $δ{Y}=0.08$ to $δ{Y}\geq0.12$, depending on the presumed helium distributions. Helium-enriched stars are comparable in number to normal stars within these clusters, and may even dominate, comprising approximately $\sim$40\% to over 80\% of the total stellar population. However, given the great distance of the Large Magellanic Cloud, it is difficult to rule out the significant impact of crowding, so our results may only represent an upper limit.

astro-ph.GA↗

arXiv2025

FaceSwapGuard: Safeguarding Facial Privacy from DeepFake Threats through Identity Obfuscation

DeepFakes pose a significant threat to our society. One representative DeepFake application is face-swapping, which replaces the identity in a facial image with that of a victim. Although existing methods partially mitigate these risks by degrading the quality of swapped images, they often fail to disrupt the identity transformation effectively. To fill this gap, we propose FaceSwapGuard (FSG), a novel black-box defense mechanism against deepfake face-swapping threats. Specifically, FSG introduces imperceptible perturbations to a user's facial image, disrupting the features extracted by identity encoders. When shared online, these perturbed images mislead face-swapping techniques, causing them to generate facial images with identities significantly different from the original user. Extensive experiments demonstrate the effectiveness of FSG against multiple face-swapping techniques, reducing the face match rate from 90\% (without defense) to below 10\%. Both qualitative and quantitative studies further confirm its ability to confuse human perception, highlighting its practical utility. Additionally, we investigate key factors that may influence FSG and evaluate its robustness against various adaptive adversaries.

cs.CR↗

arXiv2025

Monitoring Vibrational Evolution in Jahn-Teller Effect by Raman Imaging

The Jahn-Teller effect (JTE) reduces the geometrical symmetry of a system with degenerate electronic states via vibronic coupling, playing a pivotal role in molecular and condensed systems. In this Letter, we propose that vibrational resolved tip-enhanced Raman scattering images can visualize the vibrational evolutions in JTE in real space. Taking an experimentally viable single zinc phthalocyanine (ZnPc) molecule as a proof-of-principle example, not only the degenerate vibrational splitting but also the overlooked vibration mixing caused by the JTE in its anionic form can be straightforwardly characterized by Raman images. Leveraging Raman images, the controllable configuration of JTE distortion with partial isotopic substitution could be further identified. These findings establish a practical protocol to monitor the detailed vibrational evolutions when a single molecule experiences JTE, opening a door for visualization of spontaneous symmetry breaking in molecular and solid-state systems.

cond-mat.mes-hall↗

arXiv2025

Overview of the Amphion Toolkit (v0.2)

Amphion is an open-source toolkit for Audio, Music, and Speech Generation, designed to lower the entry barrier for junior researchers and engineers in these fields. It provides a versatile framework that supports a variety of generation tasks and models. In this report, we introduce Amphion v0.2, the second major release developed in 2024. This release features a 100K-hour open-source multilingual dataset, a robust data preparation pipeline, and novel models for tasks such as text-to-speech, audio coding, and voice conversion. Furthermore, the report includes multiple tutorials that guide users through the functionalities and usage of the newly released models.

cs.SD↗

arXiv2025

HCMRM: A High-Consistency Multimodal Relevance Model for Search Ads

Search advertising is essential for merchants to reach the target users on short video platforms. Short video ads aligned with user search intents are displayed through relevance matching and bid ranking mechanisms. This paper focuses on improving query-to-video relevance matching to enhance the effectiveness of ranking in ad systems. Recent vision-language pre-training models have demonstrated promise in various multimodal tasks. However, their contribution to downstream query-video relevance tasks is limited, as the alignment between the pair of visual signals and text differs from the modeling of the triplet of the query, visual signals, and video text. In addition, our previous relevance model provides limited ranking capabilities, largely due to the discrepancy between the binary cross-entropy fine-tuning objective and the ranking objective. To address these limitations, we design a high-consistency multimodal relevance model (HCMRM). It utilizes a simple yet effective method to enhance the consistency between pre-training and relevance tasks. Specifically, during the pre-training phase, along with aligning visual signals and video text, several keywords are extracted from the video text as pseudo-queries to perform the triplet relevance modeling. For the fine-tuning phase, we introduce a hierarchical softmax loss, which enables the model to learn the order within labels while maximizing the distinction between positive and negative samples. This promotes the fusion ranking of relevance and bidding in the subsequent ranking stage. The proposed method has been deployed in the Kuaishou search advertising system for over a year, contributing to a 6.1% reduction in the proportion of irrelevant ads and a 1.4% increase in ad revenue.

cs.IR↗

arXiv2025

DULRTC-RME: A Deep Unrolled Low-rank Tensor Completion Network for Radio Map Estimation

Radio maps enrich radio propagation and spectrum occupancy information, which provides fundamental support for the operation and optimization of wireless communication systems. Traditional radio maps are mainly achieved by extensive manual channel measurements, which is time-consuming and inefficient. To reduce the complexity of channel measurements, radio map estimation (RME) through novel artificial intelligence techniques has emerged to attain higher resolution radio maps from sparse measurements or few observations. However, black box problems and strong dependency on training data make learning-based methods less explainable, while model-based methods offer strong theoretical grounding but perform inferior to the learning-based methods. In this paper, we develop a deep unrolled low-rank tensor completion network (DULRTC-RME) for radio map estimation, which integrates theoretical interpretability and learning ability by unrolling the tedious low-rank tensor completion optimization into a deep network. It is the first time that algorithm unrolling technology has been used in the RME field. Experimental results demonstrate that DULRTC-RME outperforms existing RME methods.

eess.SP↗

arXiv2025

DistrEE: Distributed Early Exit of Deep Neural Network Inference on Edge Devices

Distributed DNN inference is becoming increasingly important as the demand for intelligent services at the network edge grows. By leveraging the power of distributed computing, edge devices can perform complicated and resource-hungry inference tasks previously only possible on powerful servers, enabling new applications in areas such as autonomous vehicles, industrial automation, and smart homes. However, it is challenging to achieve accurate and efficient distributed edge inference due to the fluctuating nature of the actual resources of the devices and the processing difficulty of the input data. In this work, we propose DistrEE, a distributed DNN inference framework that can exit model inference early to meet specific quality of service requirements. In particular, the framework firstly integrates model early exit and distributed inference for multi-node collaborative inferencing scenarios. Furthermore, it designs an early exit policy to control when the model inference terminates. Extensive simulation results demonstrate that DistrEE can efficiently realize efficient collaborative inference, achieving an effective trade-off between inference latency and accuracy.

cs.DC↗

arXiv2025

Exploring the latent space of diffusion models directly through singular value decomposition

Despite the groundbreaking success of diffusion models in generating high-fidelity images, their latent space remains relatively under-explored, even though it holds significant promise for enabling versatile and interpretable image editing capabilities. The complicated denoising trajectory and high dimensionality of the latent space make it extremely challenging to interpret. Existing methods mainly explore the feature space of U-Net in Diffusion Models (DMs) instead of the latent space itself. In contrast, we directly investigate the latent space via Singular Value Decomposition (SVD) and discover three useful properties that can be used to control generation results without the requirements of data collection and maintain identity fidelity generated images. Based on these properties, we propose a novel image editing framework that is capable of learning arbitrary attributes from one pair of latent codes destined by text prompts in Stable Diffusion Models. To validate our approach, extensive experiments are conducted to demonstrate its effectiveness and flexibility in image editing. We will release our codes soon to foster further research and applications in this area.

cs.CV↗

arXiv2025

Corrector estimates and homogenization errors of unsteady flow ruled by Darcy's law

Focusing on Darcy's law incorporating memory effects, this paper studies non-stationary Stokes equations on perforated domains. We establish a sharp homogenization error for both velocity and pressure in terms of the energy norm. The main challenge lies in gauging the boundary layers induced by the incompressibility condition. To address this, we construct boundary-layer correctors using Bogovskii's operator. Also, the present work provides detailed regularity estimates for these correctors, where a significant difficulty arises from the incompatibility between initial and boundary values. The methodologies developed herein hold great potential for tackling the same issue in other evolutionary models beyond a homogenization setting.

math.AP↗

arXiv2025

Least-Squares Problem Over Probability Measure Space

In this work, we investigate the variational problem $$ρ_x^\ast = \text{argmin}_{ρ_x} D(G\#ρ_x, ρ_y)\,, $$ where $D$ quantifies the difference between two probability measures, and ${G}$ is a forward operator that maps a variable $x$ to $y=G(x)$. This problem can be regarded as an analogue of its counterpart in linear spaces (e.g., Euclidean spaces), $\text{argmin}_x \|G(x) - y\|^2$. Similar to how the choice of norm $\|\cdot\|$ influences the optimizer in $\mathbb R^d$ or other linear spaces, the minimizer in the probabilistic variational problem also depends on the choice of $D$. Our findings reveal that using a $ϕ$-divergence for $D$ leads to the recovery of a conditional distribution of $ρ_y$, while employing the Wasserstein distance results in the recovery of a marginal distribution.

math.OC↗

arXiv2024

Language model driven: a PROTAC generation pipeline with dual constraints of structure and property

The imperfect modeling of ternary complexes has limited the application of computer-aided drug discovery tools in PROTAC research and development. In this study, an AI-assisted approach for PROTAC molecule design pipeline named LM-PROTAC was developed, which stands for language model driven Proteolysis Targeting Chimera, by embedding a transformer-based generative model with dual constraints on structure and properties, referred to as the DCT. This study utilized the fragmentation representation of molecules and developed a language model driven pipeline. Firstly, a language model driven affinity model for protein compounds to screen molecular fragments with high affinity for the target protein. Secondly, structural and physicochemical properties of these fragments were constrained during the generation process to meet specific scenario requirements. Finally, a two-round screening of the preliminary generated molecules using a multidimensional property prediction model to generate a batch of PROTAC molecules capable of degrading disease-relevant target proteins for validation in vitro experiments, thus achieving a complete solution for AI-assisted PROTAC drug generation. Taking the tumor key target Wnt3a as an example, the LM-PROTAC pipeline successfully generated PROTAC molecules capable of inhibiting Wnt3a. The results show that DCT can efficiently generate PROTAC that targets and hydrolyses Wnt3a.

q-bio.QM↗

arXiv2024

Taming Diffusion Prior for Image Super-Resolution with Domain Shift SDEs

Diffusion-based image super-resolution (SR) models have attracted substantial interest due to their powerful image restoration capabilities. However, prevailing diffusion models often struggle to strike an optimal balance between efficiency and performance. Typically, they either neglect to exploit the potential of existing extensive pretrained models, limiting their generative capacity, or they necessitate a dozens of forward passes starting from random noises, compromising inference efficiency. In this paper, we present DoSSR, a Domain Shift diffusion-based SR model that capitalizes on the generative powers of pretrained diffusion models while significantly enhancing efficiency by initiating the diffusion process with low-resolution (LR) images. At the core of our approach is a domain shift equation that integrates seamlessly with existing diffusion models. This integration not only improves the use of diffusion prior but also boosts inference efficiency. Moreover, we advance our method by transitioning the discrete shift process to a continuous formulation, termed as DoS-SDEs. This advancement leads to the fast and customized solvers that further enhance sampling efficiency. Empirical results demonstrate that our proposed method achieves state-of-the-art performance on synthetic and real-world datasets, while notably requiring only 5 sampling steps. Compared to previous diffusion prior based methods, our approach achieves a remarkable speedup of 5-7 times, demonstrating its superior efficiency. Code: https://github.com/QinpengCui/DoSSR.

cs.CV↗