arXiv subjects

Xuan Liu

Publications and source records attributed to Xuan Liu.

At least 73 records · Page 4Linked to original sources

arXiv2025

Spotlighter: Revisiting Prompt Tuning from a Representative Mining View

CLIP's success has demonstrated that prompt tuning can achieve robust cross-modal semantic alignment for tasks ranging from open-domain recognition to fine-grained classification. However, redundant or weakly relevant feature components introduce noise and incur unnecessary computational costs. In this work, we propose Spotlighter, a lightweight token-selection framework that simultaneously enhances accuracy and efficiency in prompt tuning. Spotlighter evaluates each visual token's activation from both sample-wise and semantic-wise perspectives and retains only the top-scoring tokens for downstream prediction. A class-specific semantic memory bank of learned prototypes refines this selection, ensuring semantic representativeness and compensating for discarded features. To further prioritize informative signals, we introduce a two-level ranking mechanism that dynamically weights token--prototype interactions. Across 11 few-shot benchmarks, Spotlighter outperforms CLIP by up to 11.19\% in harmonic mean accuracy and achieves up to 0.8K additional FPS, with only 21 extra parameters. These results establish Spotlighter as an effective and scalable baseline for prompt tuning. Code for our method will be available at https://github.com/greatest-gourmet/Spotlighter.

cs.CV↗

arXiv2025

Asymptotic universal moment matching properties of normal distributions

Moment matching is an easy-to-implement and usually effective method to reduce variance of Monte Carlo simulation estimates. On the other hand, there is no guarantee that moment matching will always reduce simulation variance for general integration problems at least asymptotically, i.e. when the number of samples is large. We study the characterization of conditions on a given underlying distribution $X$ under which asymptotic variance reduction is guaranteed for a general integration problem $\mathbb{E}[f(X)]$ when moment matching techniques are applied. We show that a sufficient and necessary condition for such asymptotic variance reduction property is $X$ being a normal distribution. Moreover, when $X$ is a normal distribution, formulae for efficient estimation of simulation variance for (first and second order) moment matching Monte Carlo are obtained. These formulae allow estimations of simulation variance as by-products of the simulation process, in a way similar to variance estimations for plain Monte Carlo. Moreover, we propose non-linear moment matching schemes for any given continuous distribution such that asymptotic variance reduction is guaranteed.

math.ST↗

arXiv2025

Continuous variable quantum communication with 40 pairs of entangled sideband

Constructing large-scale quantum resources is an important foundation for further improving the efficiency and scalability of quantum communication. Here, we present an efficient extraction and stable control scheme of 40 pairs of entangled sideband modes from the squeezed light by specially designing optical parametric oscillator. Utilizing the low-loss optical frequency comb control technology and the local cross-correlation algorithm, we model and manage the efficient separation process of the entangled sidebands modes facilitated by the optical filtering cavities, a maximum entanglement level of 6.5 dB is achieved. The feasibility of large-capacity quantum dense coding based on these entangled sideband modes is proved experimentally, which is of great significance for optimizing the utilization of quantum resources, thereby contributing to the advancement of large-capacity quantum communication networks and enabling the realization of more secure and efficient quantum communication systems.

quant-ph↗

arXiv2025

MammoTracker: Mask-Guided Lesion Tracking in Temporal Mammograms

Accurate lesion tracking in temporal mammograms is essential for monitoring breast cancer progression and facilitating early diagnosis. However, automated lesion correspondence across exams remains a challenges in computer-aided diagnosis (CAD) systems, limiting their effectiveness. We propose MammoTracker, a mask-guided lesion tracking framework that automates lesion localization across consecutively exams. Our approach follows a coarse-to-fine strategy incorporating three key modules: global search, local search, and score refinement. To support large-scale training and evaluation, we introduce a new dataset with curated prior-exam annotations for 730 mass and calcification cases from the public EMBED mammogram dataset, yielding over 20000 lesion pairs, making it the largest known resource for temporal lesion tracking in mammograms. Experimental results demonstrate that MammoTracker achieves 0.455 average overlap and 0.509 accuracy, surpassing baseline models by 8%, highlighting its potential to enhance CAD-based lesion progression analysis. Our dataset will be available at https://gitlab.oit.duke.edu/railabs/LoGroup/mammotracker.

cs.CV↗

arXiv2025

Relativistic quantum mechanics of charged vortex particles accelerated in a uniform electric field

The relativistic quantum-mechanical description of a charged Laguerre-Gauss beam accelerated in a uniform electric field has been fulfilled. Stationary wave eigenfunctions are rigorously derived. The evolution of the beam parameters during acceleration is considered in detail. The practically important effect of extraordinary suppression of transverse spreading of the beam is discovered, carefully analyzed, and properly explained. Our results provide direct evidence that vortex particle beams can be accelerated without destroying their intrinsic vortex properties, paving the way for high-energy vortex beam applications.

physics.acc-ph↗

arXiv2025

Generalized Gouy Rotation of Electron Vortex beams in uniform magnetic fields

The intrinsic rotation of electron vortex beams, governed by their phase structure, has been experimentally observed in magnetic fields by breaking the beam's cylindrical symmetry. However, conventional Landau states, which predict three fixed angular frequencies, cannot fully account for the existing experimental observations. To address this limitation, we introduce and derive the generalized Gouy rotation angle, which links the Gouy phase of an extended Landau state -- featuring a periodically oscillating beam width -- to the experimentally observed angular variation. In particular, this framework predicts a broader spectrum of angular frequencies and captures the reversal of rotation direction observed in electron vortex beams with negative topological charge. Calculations based on experimental parameters show good agreement with previously published data and are further validated here by numerical simulations using the Chebyshev method. Our results are, in principle, applicable to any system involving electron vortex beams in uniform magnetic fields, and provide a foundation for exploring vortex electrons in Glaser and other nonuniform magnetic fields.

quant-ph↗

arXiv2025

Energy-Embedded Neural Solvers for One-Dimensional Quantum Systems

Physics-informed neural networks (PINN) have been widely used in computational physics to solve partial differential equations (PDEs). In this study, we propose an energy-embedding-based physics-informed neural network method for solving the one-dimensional time-independent Schrödinger equation to obtain ground- and excited-state wave functions, as well as energy eigenvalues by incorporating an embedding layer to generate process-driven data. The method demonstrates high accuracy for several well-known potentials, such as the infinite potential well, harmonic oscillator potential, Woods-Saxon potential, and double-well potential. Further validation shows that the method also performs well in solving the radial Coulomb potential equation, showcasing its adaptability and extensibility. The proposed approach can be extended to solve other partial differential equations beyond the Schrödinger equation and holds promise for applications in high-dimensional quantum systems.

physics.comp-ph↗

arXiv2025

RTime-QA: A Benchmark for Atomic Temporal Event Understanding in Large Multi-modal Models

Understanding accurate atomic temporal event is essential for video comprehension. However, current video-language benchmarks often fall short to evaluate Large Multi-modal Models' (LMMs) temporal event understanding capabilities, as they can be effectively addressed using image-language models. In this paper, we introduce RTime-QA, a novel benchmark specifically designed to assess the atomic temporal event understanding ability of LMMs. RTime-QA comprises 822 high-quality, carefully-curated video-text questions, each meticulously annotated by human experts. Each question features a video depicting an atomic temporal event, paired with both correct answers and temporal negative descriptions, specifically designed to evaluate temporal understanding. To advance LMMs' temporal event understanding ability, we further introduce RTime-IT, a 14k instruction-tuning dataset that employs a similar annotation process as RTime-QA. Extensive experimental analysis demonstrates that RTime-QA presents a significant challenge for LMMs: the state-of-the-art model Qwen2-VL achieves only 34.6 on strict-ACC metric, substantially lagging behind human performance. Furthermore, our experiments reveal that RTime-IT effectively enhance LMMs' capacity in temporal understanding. By fine-tuning on RTime-IT, our Qwen2-VL achieves 65.9 on RTime-QA.

cs.CV↗

arXiv2025

The defocusing energy-supercritical inhomogeneous NLS in four space dimension

In this paper, we investigate the global well-posedness and scattering theory for the defocusing energy supcritical inhomogeneous nonlinear Schrödinger equation $iu_t + Δu =|x|^{-b} |u|^αu$ in four space dimension, where $s_c := 2- \frac{2-b}α \in (1, 2)$ and $0<b<\min \{ (s_c-1)^2+1,3-s_c\}$. We prove that if the solution has a prior bound in the critical Sobolev space, that is, $u \in L_t^\infty(I; \dot{H}_x^{s_c}(\mathbb{R}^4))$, then $u$ is global and scatters. The proof of the main results is based on the concentration-compactness/rigidity framework developed by Kenig and Merle [Invent. Math. 166 (2006)], together with a long-time Strichartz estimate, a spatially localized Morawetz estimate, and a frequency-localized Morawetz estimate.

math.AP↗

arXiv2025

LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning

In continual learning (CL), catastrophic forgetting often arises due to feature drift. This challenge is particularly prominent in the exemplar-free continual learning (EFCL) setting, where samples from previous tasks cannot be retained, making it difficult to preserve prior knowledge. To address this issue, some EFCL methods aim to identify feature spaces that minimize the impact on previous tasks while accommodating new ones. However, they rely on static features or outdated statistics stored from old tasks, which prevents them from capturing the dynamic evolution of the feature space in CL, leading to performance degradation over time. In this paper, we introduce the Drift-Resistant Space (DRS), which effectively handles feature drifts without requiring explicit feature modeling or the storage of previous tasks. A novel parameter-efficient fine-tuning approach called Low-Rank Adaptation Subtraction (LoRA-) is proposed to develop the DRS. This method subtracts the LoRA weights of old tasks from the initial pre-trained weight before processing new task data to establish the DRS for model training. Therefore, LoRA- enhances stability, improves efficiency, and simplifies implementation. Furthermore, stabilizing feature drifts allows for better plasticity by learning with a triplet loss. Our method consistently achieves state-of-the-art results, especially for long task sequences, across multiple datasets.

cs.LG↗

arXiv2025

Exploring Prosocial Irrationality for LLM Agents: A Social Cognition View

Large language models (LLMs) have been shown to face hallucination issues due to the data they trained on often containing human bias; whether this is reflected in the decision-making process of LLM Agents remains under-explored. As LLM Agents are increasingly employed in intricate social environments, a pressing and natural question emerges: Can we utilize LLM Agents' systematic hallucinations to mirror human cognitive biases, thus exhibiting irrational social intelligence? In this paper, we probe the irrational behavior among contemporary LLM Agents by melding practical social science experiments with theoretical insights. Specifically, We propose CogMir, an open-ended Multi-LLM Agents framework that utilizes hallucination properties to assess and enhance LLM Agents' social intelligence through cognitive biases. Experimental results on CogMir subsets show that LLM Agents and humans exhibit high consistency in irrational and prosocial decision-making under uncertain conditions, underscoring the prosociality of LLM Agents as social entities and highlighting the significance of hallucination properties. Additionally, the CogMir framework demonstrates its potential as a valuable platform for encouraging more research into the social intelligence of LLM Agents.

cs.CY↗

arXiv2025

Splitting finite element approximations for quasi-static electroporoelasticity equations

The electroporoelasticity model, which couples Maxwell's equations with Biot's equations, plays a critical role in applications such as water conservancy exploration, earthquake early warning, and various other fields. This work focuses on investigating its well-posedness and analyzing error estimates for a splitting backward Euler finite element method. We first define a weak solution consistent with the finite element framework. Then, we prove the uniqueness and existence of such a solution using the Galerkin method and derive a priori estimates for high-order regularity. Using a splitting technique, we define an approximate splitting solution and analyze its convergence order. Next, we apply Nedelec's curl-conforming finite elements, Lagrange elements, and the backward Euler method to construct a fully discretized scheme. We demonstrate the stability of the splitting numerical solution and provide error estimates for its convergence order in both temporal and spatial variables. Finally, we present numerical experiments to validate the theoretical results, showing that our method significantly reduces computational complexity compared to the classical finite element method.

math.NA↗

arXiv2025

Multi-Class Segmentation of Aortic Branches and Zones in Computed Tomography Angiography: The AortaSeg24 Challenge

Multi-class segmentation of the aorta in computed tomography angiography (CTA) scans is essential for diagnosing and planning complex endovascular treatments for patients with aortic dissections. However, existing methods reduce aortic segmentation to a binary problem, limiting their ability to measure diameters across different branches and zones. Furthermore, no open-source dataset is currently available to support the development of multi-class aortic segmentation methods. To address this gap, we organized the AortaSeg24 MICCAI Challenge, introducing the first dataset of 100 CTA volumes annotated for 23 clinically relevant aortic branches and zones. This dataset was designed to facilitate both model development and validation. The challenge attracted 121 teams worldwide, with participants leveraging state-of-the-art frameworks such as nnU-Net and exploring novel techniques, including cascaded models, data augmentation strategies, and custom loss functions. We evaluated the submitted algorithms using the Dice Similarity Coefficient (DSC) and Normalized Surface Distance (NSD), highlighting the approaches adopted by the top five performing teams. This paper presents the challenge design, dataset details, evaluation metrics, and an in-depth analysis of the top-performing algorithms. The annotated dataset, evaluation code, and implementations of the leading methods are publicly available to support further research. All resources can be accessed at https://aortaseg24.grand-challenge.org.

eess.IV↗

arXiv2025

Holographic Schwinger effect in flavor-dependent systems

The holographic Schwinger effect is investigated in systems with $N_{f}=0$, $N_{f}=2$, and $N_{f}=2+1$ using the Einstein-Maxwell-dilaton (EMD) model, incorporating equation of state and baryon number susceptibility information from lattice quantum chromodynamics (QCD). It is found that the critical electric field is smallest for $N_{f}=0$, indicating that the Schwinger effect is more likely to occur than in systems with $N_{f}=2$ and $N_{f}=2+1$. The critical electric field decreases with increasing chemical potential and temperature across all systems. Additionally, potential analysis confirms that the maximum total potential energy increases with the number of flavors, suggesting that existing particles may reduce the probability of particle pair production.

hep-ph↗

arXiv2025

Scattering theory for the defocusing 3d NLS in the exterior of a strictly convex obstacle

In this paper, we investigate the global well-posedness and scattering theory for the defocusing nonlinear Schrödinger equation $iu_t + Δ_Ωu = |u|^αu$ in the exterior domain $Ω$ of a smooth, compact and strictly convex obstacle in $\mathbb{R}^3$. It is conjectured that in Euclidean space, if the solution has a prior bound in the critical Sobolev space, that is, $u \in L_t^\infty(I; \dot{H}_x^{s_c}(\mathbb{R}^3))$ with $s_c := \frac{3}{2} - \frac{2}α \in (0, \frac{3}{2})$, then $u$ is global and scatters. In this paper, assuming that this conjecture holds, we prove that if $u$ is a solution to the nonlinear Schrödinger equation in exterior domain $Ω$ with Dirichlet boundary condition and satisfies $u \in L_t^\infty(I; \dot{H}^{s_c}_D(Ω))$ with $s_c \in \left[\frac{1}{2}, \frac{3}{2}\right)$, then $u$ is global and scatters. The proof of the main results relies on the concentration-compactness/rigidity argument of Kenig and Merle [Invent. Math. {\bf 166} (2006)]. The main difficulty is to construct minimal counterexamples when the scaling and translation invariance breakdown on $Ω$. To achieve this, two key ingredients are required. First, we adopt the approach of Killip, Visan, and Zhang [Amer. J. Math. {\bf 138} (2016)] to derive the linear profile decomposition for the linear propagator $e^{itΔ_Ω}$ in $\dot{H}^{s_c}(Ω)$. The second ingredient is the embedding of the nonlinear profiles. More precisely, we need to demonstrate that nonlinear solutions in the limiting geometries, which exhibit global spacetime bounds, can be embedded back into $Ω$. Finally, to rule out the minimal counterexamples, we will establish long-time Strichartz estimates for the exterior domain NLS, along with spatially localized and frequency-localized Morawetz estimates.

math.AP↗

arXiv2025

Mjolnir: Breaking the Shield of Perturbation-Protected Gradients via Adaptive Diffusion

Perturbation-based mechanisms, such as differential privacy, mitigate gradient leakage attacks by introducing noise into the gradients, thereby preventing attackers from reconstructing clients' private data from the leaked gradients. However, can gradient perturbation protection mechanisms truly defend against all gradient leakage attacks? In this paper, we present the first attempt to break the shield of gradient perturbation protection in Federated Learning for the extraction of private information. We focus on common noise distributions, specifically Gaussian and Laplace, and apply our approach to DNN and CNN models. We introduce Mjolnir, a perturbation-resilient gradient leakage attack that is capable of removing perturbations from gradients without requiring additional access to the original model structure or external data. Specifically, we leverage the inherent diffusion properties of gradient perturbation protection to develop a novel diffusion-based gradient denoising model for Mjolnir. By constructing a surrogate client model that captures the structure of perturbed gradients, we obtain crucial gradient data for training the diffusion model. We further utilize the insight that monitoring disturbance levels during the reverse diffusion process can enhance gradient denoising capabilities, allowing Mjolnir to generate gradients that closely approximate the original, unperturbed versions through adaptive sampling steps. Extensive experiments demonstrate that Mjolnir effectively recovers the protected gradients and exposes the Federated Learning process to the threat of gradient leakage, achieving superior performance in gradient denoising and private data recovery.

cs.LG↗

arXiv2024

Global well-posedness and scattering for mass-critical inhomogeneous NLS when $d\ge3$

We prove global well-posedness and scattering for solutions to the mass-critical inhomogeneous nonlinear Schrödinger equation $i\partial_{t}u+Δu=\pm |x|^{-b}|u|^{\frac{4-2b}{d}}u$ for large $L^2(\mathbb{R} ^d)$ initial data with $d\ge3,0 0$) are that the presence of the inhomogeneity $|x|^{-b}$ creates a nontrivial singularity at the origin, and breaks the translation symmetry as well as the Galilean invariance of the equation, which makes the establishment of the profile decomposition and long time Strichartz estimates more difficult. To overcome these difficulties, we perform the concentration compactness/rigidity methods of [Kenig and Merle, Invent. Math. (2006)] in the Lorentz space framework, and reduces the problem to the exclusion of almost periodic solutions. The exclusion of these solutions will utilize fractional estimates and long time Strichartz estimates in Lorentz spaces. In our study, we obseve that the decay of the inhomogeneity $|x|^{-b}$ at infinity prevents the concentration of the almost periodic solution at infinity in either physical or frequency space. Therefore, we can use classical Morawetz estimates, rather than interaction Morawetz estimates, to exclude the existence of the quasi-soliton.

math.AP↗

arXiv2024

The Product of linear forms over function fields

The aim of this paper is to study the product of $n$ linear forms over function fields. We calculate the maximum value of the minima of the forms with determinant one when $n$ is small. The value is equal to the natural bound given by algebraic number theory. Our proof is based on a reduction theory of diagonal group orbits on homogeneous spaces. We also show that the forms defined algebraically correspond to periodic orbits with respect to the diagonal group actions.

math.NT↗