arXiv subjects

Wei Hao

Publications and source records attributed to Wei Hao.

At least 19 recordsLinked to original sources

arXiv2026

SpecF2M: A Spectral-Aware Multi-task Network Estimating Axial Length and Refractive Error from Pediatric Fundus Photographs

Spherical Equivalent Refraction (SER) and Axial Length (AL) are core indicators for pediatric myopia screening, yet their measurements require dedicated biometry and cycloplegic refraction. Fundus photography offers an accessible imaging modality, as myopia-related posterior-pole changes are visible in 45$^\circ$ fundus images. However, these cues are often low-contrast, spatially diffuse, and multi-scale. Moreover, AL, Sphere (SPH), and Cylinder (CYL) share partially overlapping but non-identical anatomical correlates. We propose SpecF2M, a spectral-aware multi-task network for estimating AL and SER components from pediatric fundus photographs. SpecF2M integrates a deterministic anatomy-guided enhancement module, a hybrid spatial--spectral backbone combining MixCNN and Hybrid Spectral Learning (HSL) blocks, and an expert-routing head for component-level estimation of AL, SPH, and CYL. On a pediatric cohort of 4,359 eligible child visits and 6,966 fundus images, SpecF2M outperforms controlled CNN/ViT baselines for AL and SPH estimation, achieving MAEs of 0.5347 mm and 0.7062 D, respectively. Component-level analysis further reveals asymmetric task coupling, where CYL exhibits weaker association with fundus-derived myopic patterns than AL/SPH. These results support fundus-based, screening-oriented estimation of pediatric myopia indicators, while external validation remains necessary before deployment.

eess.IV↗

arXiv2026

Tuning Cu/Diamond Interfacial Thermal Conductance via Nitrogen-Termination Engineering

Cu-diamond composites are recognized as promising high-thermal-conductivity candidates for electronic cooling, offering tunable properties and competitive cost. However, their performance is significantly limited by the poor Cu/diamond interfacial thermal conductance (ITC). Here, we propose a nitrogen-termination strategy to tune the ITC of Cu/diamond interfaces and unravel atomistic mechanisms by which nitride interlayers tailor phonon transport. Based on the MACE machine-learning interatomic potential (MLIP) framework, we fine-tune the pre-trained MACE-MPA-0 foundation model by incorporating customized C-N-Cu training datasets. Through MLIP-driven lattice dynamics simulations, we demonstrate that an atomically flat N-termination on diamond enhances the ITC by 21% compared to the bare Cu/diamond interface. Mode-resolved phonon spectroscopy reveals that the LA phonons with frequency above 4 THz and wavevectors near {\Gamma}-X and {\Gamma}-U directions are selectively modulated by N-termination engineering. Analyses of local vibrational states and interfacial bonding further indicate that the N-termination on diamond tunes the interfacial heat conduction via surficial mass modification and bonding regulation, as evidenced by variations in LDOS overlap and COHP spectra. These findings open venues for tuning heat transfer across Cu/diamond interfaces via non-metallic modification, which avoids the graphitization issues associated with metallic coatings, and provide novel guidelines for upgrading the phonon-mediated heat transfer in Cu-diamond composites.

physics.comp-ph↗

arXiv2025

Copula Structural Equation Models for Mediation Pathway Analysis

Structural equation models (SEMs) are fundamental to causal mediation pathway discovery. However, traditional SEM approaches often rely on \emph{ad hoc} model specifications when handling complex data structures such as mixed data types or non-normal data in which Gaussian assumptions for errors are rather restrictive. The invocation of copula dependence modeling methods to extend the classical linear SEMs mitigates several of key technical limitations, offering greater modeling flexibility to analyze non-Gaussian data. This paper presents a selective review of major developments in this area, highlighting recent advancements and their methodological implications.

stat.ME↗

arXiv2025

Strangeonium spectrum with the screening effects and interpretation of $h_1(1911)$ and $X(2300)$ observed by BESIII

Motivated by two news states $h_1(1911)$ and $X(2300)$ observed by BESIII, we have investigated the mass spectrum and the strong decay properties of the strangeonium mesons within the modified Godfrey-Isgur model by considering the screening effects. We have determined the free parameters using the masses and widths of the well established $s\bar{s}$ states $ϕ(1020)$, $ϕ(1680)$, $h_1(1415)$, $f_2^\prime(1525)$, and $ϕ_3(1850)$. According to our results, $h_1(1911)$ and $X(2300)$ could be well explained as states $h_1(2^1P_1)$ and $h_1(3^1P_1)$ $s\bar{s}$ states, respectively. Meanwhile, the possible assignments of $X(2000)$, $η_2(1870)$, and $ϕ(2170)$ as $3^3S_1$, $1^1D_2$, and $2^3D_1$ are also discussed. Furthermore, the masses and widths of the $2S$, $3S$, $1P$, $2P$, $3P$, $1D$, and $2D$ $s\bar{s}$ states are also given and compared with various theoretical predictions, which is helpful for the observations and confirmations of these states in future.

hep-ph↗

arXiv2025

User Trajectory Prediction Unifying Global and Local Temporal Information

Trajectory prediction is essential for formulating proactive strategies that anticipate user mobility and support advance preparation. Therefore, how to reduce the forecasting error in user trajectory prediction within an acceptable inference time arises as an interesting issue. However, trajectory data contains both global and local temporal information, complicating the extraction of the complete temporal pattern. Moreover, user behavior occurs over different time scales, increasing the difficulty of capturing behavioral patterns. To address these challenges, a trajectory prediction model based on multilayer perceptron (MLP), multi-scale convolutional neural network (MSCNN), and cross-attention (CA) is proposed. Specifically, MLP is used to extract the global temporal information of each feature. In parallel, MSCNN is employed to extract the local temporal information by modeling interactions among features within a local temporal range. Convolutional kernels with different sizes are used in MSCNN to capture temporal information at multiple resolutions, enhancing the model's adaptability to different behavioral patterns. Finally, CA is applied to fuse the global and local temporal information. Experimental results show that our model reduces mean squared error (MSE) by 5.04% and mean absolute error (MAE) by 4.35% compared with ModernTCN in 12-step prediction, while maintaining similar inference time.

cs.LG↗

arXiv2025

The Study of Pole Trajectory within a bare state in the coupled channel model

We investigate two-particle scattering and two-particle scattering with a bare basis state using Hamiltonian Effective Field Theory (HEFT). We analyze the distribution of two-body scattering poles in the momentum and energy planes under relativistic conditions. Compared to the non-relativistic case, there are significant differences in the distribution of bound state poles and resonance poles in the relativistic case, primarily due to the square root term in the relativistic formula. By considering pure two-particle scattering, we examine the relationship between the form factor and the number of poles. Additionally, we clearly elucidate the effects of attractive and repulsive interactions on the bound state poles and resonance poles. More importantly, we extend our model by including a bare state and explore the poles originating from the bare state or coupled channels through the trajectories of pole positions, as well as the compositeness of bound states.

hep-ph↗

arXiv2025

Diversity Helps Jailbreak Large Language Models

We have uncovered a powerful jailbreak technique that leverages large language models' ability to diverge from prior context, enabling them to bypass safety constraints and generate harmful outputs. By simply instructing the LLM to deviate and obfuscate previous attacks, our method dramatically outperforms existing approaches, achieving up to a 62.83% higher success rate in compromising ten leading chatbots, including GPT-4, Gemini, and Llama, while using only 12.9% of the queries. This revelation exposes a critical flaw in current LLM safety training, suggesting that existing methods may merely mask vulnerabilities rather than eliminate them. Our findings sound an urgent alarm for the need to revolutionize testing methodologies to ensure robust and reliable LLM security.

cs.CL↗

arXiv2025

Bottomonium meson spectrum with quenched and unquenched quark models

An open question in hadronic phenomenology concerns the ``unquenching" effects of higher Fock space components on the leading Fock space description of hadrons. We address this by making a comparison of the bottomonium spectrum as computed with the relativized Godfrey-Isgur quark model and an unquenched coupled channel model driven by the ``$^3P_0$" mechanism of hadronic decay. Our results show that both models can describe the spectrum well, indicating that the influence of coupled channel effects can be largely absorbed into the parameters of the quenched quark model. This conclusion is reinforced by a perturbative calculation that shows that the spin-dependence of mass splittings due to mixing with the continuum recapitulates quenched quark model spin-dependent interactions. We also show that softening of the quark-antiquark wavefunction due to continuum mixing improves the description of vector bottomonium decay constants. Together, these results illustrate and substantiate the surprising robustness of simple constituent quark model descriptions of hadrons.

hep-ph↗

arXiv2025

Learning to Rewrite: Generalized LLM-Generated Text Detection

Large language models (LLMs) present significant risks when used to generate non-factual content and spread disinformation at scale. Detecting such LLM-generated content is crucial, yet current detectors often struggle to generalize in open-world contexts. We introduce Learning2Rewrite, a novel framework for detecting AI-generated text with exceptional generalization to unseen domains. Our method leverages the insight that LLMs inherently modify AI-generated content less than human-written text when tasked with rewriting. By training LLMs to minimize alterations on AI-generated inputs, we amplify this disparity, yielding a more distinguishable and generalizable edit distance across diverse text distributions. Extensive experiments on data from 21 independent domains and four major LLMs (GPT-3.5, GPT-4, Gemini, and Llama-3) demonstrate that our detector outperforms state-of-the-art detection methods by up to 23.04% in AUROC for in-distribution tests, 37.26% for out-of-distribution tests, and 48.66% under adversarial attacks. Our unique training objective ensures better generalizability compared to directly training for classification, when leveraging the same amount of parameters. Our findings suggest that reinforcing LLMs' inherent rewriting tendencies offers a robust and scalable solution for detecting AI-generated text.

cs.CL↗

arXiv2025

Coupled channel effects for the bottom-strange mesons

We have calculated the mass spectrum of $B_s$ mesons within a nonrelativistic potential model considering coupled channel effects, and the corresponding strong decay widths within the $^3P_0$ model using the numerically calculated wave functions. By comparing with the available experimental data, we find that the states $B_s$, $B_s^*$, $B_{s1}(5830)$, and $B_{s2}^*(5840)$ could be interpreted as the $B_s(1^1S_0)$, $B_s(1^3S_1)$, $B_s(1P^\prime)$, and $B_s(1^3P_2)$, respectively. Although the quantum numbers of the newly observed $B_s(6064)$ and $B_s(6158)$ states have not been determined, our results support the assignments of $B_s(1^3D_3)$ and $B_s(1^3D_1)$ for them. Our predictions are helpful in searching for the bottom-strange meson in future experiments.

hep-ph↗

arXiv2024

Spectrum and decay properties of the charmed mesons involving the coupled channel effects

The mass spectrum of the charmed mesons is investigated by considering the coupled channel effects within the nonrelativistic potential model. The predicted masses of the charmed mesons are in agreement with experimental data. The strong decay properties are further analyzed within the $^3P_0$ model by using numerical wave functions obtained from nonrelativistic potential model. Based on the predicted masses and decay properties, we give a classification of the recently observed charmed states. Especially, we have effectively explained the masses and decay properties of the $D_1^*(2600)$ and $D_1^*(2760)$ by considering the $S$-$D$ mixing. Furthermore, the predicted masses and decay properties of the $2P$ wave states are helpful to search for them experimentally in future.

hep-ph↗

arXiv2024

I Can Hear You: Selective Robust Training for Deepfake Audio Detection

Recent advances in AI-generated voices have intensified the challenge of detecting deepfake audio, posing risks for scams and the spread of disinformation. To tackle this issue, we establish the largest public voice dataset to date, named DeepFakeVox-HQ, comprising 1.3 million samples, including 270,000 high-quality deepfake samples from 14 diverse sources. Despite previously reported high accuracy, existing deepfake voice detectors struggle with our diversely collected dataset, and their detection success rates drop even further under realistic corruptions and adversarial attacks. We conduct a holistic investigation into factors that enhance model robustness and show that incorporating a diversified set of voice augmentations is beneficial. Moreover, we find that the best detection models often rely on high-frequency features, which are imperceptible to humans and can be easily manipulated by an attacker. To address this, we propose the F-SAT: Frequency-Selective Adversarial Training method focusing on high-frequency components. Empirical results demonstrate that using our training dataset boosts baseline model performance (without robust training) by 33%, and our robust training further improves accuracy by 7.7% on clean samples and by 29.3% on corrupted and attacked samples, over the state-of-the-art RawNet3 model.

cs.SD↗

arXiv2024

Towards Consistent Object Detection via LiDAR-Camera Synergy

As human-machine interaction continues to evolve, the capacity for environmental perception is becoming increasingly crucial. Integrating the two most common types of sensory data, images, and point clouds, can enhance detection accuracy. Currently, there is no existing model capable of detecting an object's position in both point clouds and images while also determining their corresponding relationship. This information is invaluable for human-machine interactions, offering new possibilities for their enhancement. In light of this, this paper introduces an end-to-end Consistency Object Detection (COD) algorithm framework that requires only a single forward inference to simultaneously obtain an object's position in both point clouds and images and establish their correlation. Furthermore, to assess the accuracy of the object correlation between point clouds and images, this paper proposes a new evaluation metric, Consistency Precision (CP). To verify the effectiveness of the proposed framework, an extensive set of experiments has been conducted on the KITTI and DAIR-V2X datasets. The study also explored how the proposed consistency detection method performs on images when the calibration parameters between images and point clouds are disturbed, compared to existing post-processing methods. The experimental results demonstrate that the proposed method exhibits excellent detection performance and robustness, achieving end-to-end consistency detection. The source code will be made publicly available at https://github.com/xifen523/COD.

cs.CV↗

arXiv2024

NTIRE 2024 Challenge on Image Super-Resolution (x4): Methods and Results

This paper reviews the NTIRE 2024 challenge on image super-resolution ($\times$4), highlighting the solutions proposed and the outcomes obtained. The challenge involves generating corresponding high-resolution (HR) images, magnified by a factor of four, from low-resolution (LR) inputs using prior information. The LR images originate from bicubic downsampling degradation. The aim of the challenge is to obtain designs/solutions with the most advanced SR performance, with no constraints on computational resources (e.g., model size and FLOPs) or training data. The track of this challenge assesses performance with the PSNR metric on the DIV2K testing dataset. The competition attracted 199 registrants, with 20 teams submitting valid entries. This collective endeavour not only pushes the boundaries of performance in single-image SR but also offers a comprehensive overview of current trends in this field.

cs.CV↗

arXiv2024

Beauty-charm Meson Family with Coupled Channel Effects and Their Strong Decays

We systematically study the mass spectra and their two-body hadronic decays of the beauty-charm meson family considering the coupled channel effects. Our results can good explain the observed $B_c$ meson spectrum and the prediction of the mass spectrum for unobserved beauty-charm mesons can be tested in future experiments. For the coupled channel components, we predicted the $1S$ state in beauty-charm meson family is about $4\%$, while the $2S$, $1P$, $2P$, $1D$, and $2D$ states are about $14\%$, $10\%$, $33\%$, and $17\%$ respectively. For the $3S$, $2P$ and $2D$ states, the strong decay is allowed, The two-body hadronic decay widths of the $3^1S_0$, $3^3S_1$, $2^3P_2$ states are about 110 MeV, 69 MeV, and 3 MeV, respectively. While the two-body decay widths of the $2^3D_1$, $2D$, $2D^\prime$, and $2^3D_2$ states are 60 MeV, 149 MeV, 65 MeV, and 72 MeV, respectively.

hep-ph↗

arXiv2024

UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation

Multi-object tracking (MOT) in video sequences remains a challenging task, especially in scenarios with significant camera movements. This is because targets can drift considerably on the image plane, leading to erroneous tracking outcomes. Addressing such challenges typically requires supplementary appearance cues or Camera Motion Compensation (CMC). While these strategies are effective, they also introduce a considerable computational burden, posing challenges for real-time MOT. In response to this, we introduce UCMCTrack, a novel motion model-based tracker robust to camera movements. Unlike conventional CMC that computes compensation parameters frame-by-frame, UCMCTrack consistently applies the same compensation parameters throughout a video sequence. It employs a Kalman filter on the ground plane and introduces the Mapped Mahalanobis Distance (MMD) as an alternative to the traditional Intersection over Union (IoU) distance measure. By leveraging projected probability distributions on the ground plane, our approach efficiently captures motion patterns and adeptly manages uncertainties introduced by homography projections. Remarkably, UCMCTrack, relying solely on motion cues, achieves state-of-the-art performance across a variety of challenging datasets, including MOT17, MOT20, DanceTrack and KITTI. More details and code are available at https://github.com/corfyi/UCMCTrack

cs.CV↗

arXiv2023

Analysis of Kozai Cycles in Equal-Mass Hierarchical Triple Supermassive Black Hole Mergers in the Presence of a Stellar Cluster

Supermassive black holes (SMBHs) play an important role in galaxy evolution. Binary and triple SMBHs can form after galaxy mergers. A third SMBH may accelerate the SMBH merging process, possibly through the Kozai mechanism. We use N -body simulations to analyze oscillations in the orbital elements of hierarchical triple SMBHs with surrounding star clusters in galaxy centers. We find that SMBH triples spend only a small fraction of time in the hierarchical merger phase (i.e., a binary SMBH with a distant third SMBH perturber). Most of the time, the enclosed stellar mass within the orbits of the innermost or the outermost SMBH is comparable to the SMBH masses, indicating that the influence of the surrounding stellar population cannot be ignored. We search for Eccentric Kozai-Lidov (EKL) oscillations for which (i) the eccentricity of the inner binary and inclination are both oscillate and are anti-phase or in-phase and (ii) the oscillation period is consistent with EKL timescale. We find that EKL oscillations are short-lived and rare: the triple SMBH spends around 3% of its time in this phase over the ensemble of simulations, reaching around 8% in the best-case scenario. This suggests that the role of the EKL mechanism in accelerating the SMBH merger process may have been overestimated in previous studies. We follow-up with three-body simulations, using initial conditions extracted from the simulation, and the result can to some extent repeat the observed EKL-like oscillations. This comparison provides clues about why those EKL oscillations with perturbing stars are short-lived.

astro-ph.GA↗

arXiv2023

The mass spectrum and strong decay properties of the charmed-strange mesons within Godfrey-Isgur model considering the coupled-channel effects

Motivated by the recently observed $D_{s0}(2590)$ state by LHCb, we investigate the mass spectrum and the strong decay properties of the charmed-strange mesons within Godfrey-Isgur model considering the coupled-channel effects. Our results support that $D_{s0}^*(2317)$ and $D_{s1}(2460)$ can be interpreted as the $D_{s}(1^3P_0)$ and $D_{s}(1^3P_1)$ states with larger $DK$ and $D^*K$ components, respectively, and $D_{s1}(2700)$, $D_{s1}(2536)$, $D^*_{s2}(2573)$, $D_{s1}^*(2860)$, $D_{s3}^*(2860)$, and $D_{sJ}^*(3040)$ can be well interpreted as the $D_s(2^3S_1)$, $D_s(1^1P_1)$, $D_s(1^3P_2)$, $D_s(1^3D_1)$, $D_s(1^3D_3)$, and $D_s(2^1P_1)$ states, respectively. Although, $D_{s0}(2590)$ mass is about 50 MeV less than our prediction for the $D_{s}(2^1S_0)$ state, its width is still in good agreement with the one of $D_{s}(2^1S_0)$. Therefore, $D_{s0}(2590)$ state needs to be further confirmed by the experimental measurements, and the more precise information about $D_{s0}(2590)$ will shed light on its assignment of $D_{s}(2^1S_0)$. Furthermore, we predict the masses and the strong decay properties of the charmed-strange mesons with masses around 3 GeV, which would be helpful to experimentally search for these states.

hep-ph↗