arXiv subjects

Wenxuan Wang

Publications and source records attributed to Wenxuan Wang.

At least 127 records · Page 7Linked to original sources

arXiv2025

Image Difference Grounding with Natural Language

Visual grounding (VG) typically focuses on locating regions of interest within an image using natural language, and most existing VG methods are limited to single-image interpretations. This limits their applicability in real-world scenarios like automatic surveillance, where detecting subtle but meaningful visual differences across multiple images is crucial. Besides, previous work on image difference understanding (IDU) has either focused on detecting all change regions without cross-modal text guidance, or on providing coarse-grained descriptions of differences. Therefore, to push towards finer-grained vision-language perception, we propose Image Difference Grounding (IDG), a task designed to precisely localize visual differences based on user instructions. We introduce DiffGround, a large-scale and high-quality dataset for IDG, containing image pairs with diverse visual variations along with instructions querying fine-grained differences. Besides, we present a baseline model for IDG, DiffTracker, which effectively integrates feature differential enhancement and common suppression to precisely locate differences. Experiments on the DiffGround dataset highlight the importance of our IDG dataset in enabling finer-grained IDU. To foster future research, both DiffGround data and DiffTracker model will be publicly released.

cs.CV↗

arXiv2025

Towards Unified Referring Expression Segmentation Across Omni-Level Visual Target Granularities

Referring expression segmentation (RES) aims at segmenting the entities' masks that match the descriptive language expression. While traditional RES methods primarily address object-level grounding, real-world scenarios demand a more versatile framework that can handle multiple levels of target granularity, such as multi-object, single object or part-level references. This introduces great challenges due to the diverse and nuanced ways users describe targets. However, existing datasets and models mainly focus on designing grounding specialists for object-level target localization, lacking the necessary data resources and unified frameworks for the more practical multi-grained RES. In this paper, we take a step further towards visual granularity unified RES task. To overcome the limitation of data scarcity, we introduce a new multi-granularity referring expression segmentation (MRES) task, alongside the RefCOCOm benchmark, which includes part-level annotations for advancing finer-grained visual understanding. In addition, we create MRES-32M, the largest visual grounding dataset, comprising over 32.2M masks and captions across 1M images, specifically designed for part-level vision-language grounding. To tackle the challenges of multi-granularity RES, we propose UniRES++, a unified multimodal large language model that integrates object-level and part-level RES tasks. UniRES++ incorporates targeted designs for fine-grained visual feature exploration. With the joint model architecture and parameters, UniRES++ achieves state-of-the-art performance across multiple benchmarks, including RefCOCOm for MRES, gRefCOCO for generalized RES, and RefCOCO, RefCOCO+, RefCOCOg for classic RES. To foster future research into multi-grained visual grounding, our RefCOCOm benchmark, MRES-32M dataset and model UniRES++ will be publicly available at https://github.com/Rubics-Xuan/MRES.

cs.CV↗

arXiv2025

Improving the Transferability of Adversarial Attacks on Face Recognition with Diverse Parameters Augmentation

Face Recognition (FR) models are vulnerable to adversarial examples that subtly manipulate benign face images, underscoring the urgent need to improve the transferability of adversarial attacks in order to expose the blind spots of these systems. Existing adversarial attack methods often overlook the potential benefits of augmenting the surrogate model with diverse initializations, which limits the transferability of the generated adversarial examples. To address this gap, we propose a novel method called Diverse Parameters Augmentation (DPA) attack method, which enhances surrogate models by incorporating diverse parameter initializations, resulting in a broader and more diverse set of surrogate models. Specifically, DPA consists of two key stages: Diverse Parameters Optimization (DPO) and Hard Model Aggregation (HMA). In the DPO stage, we initialize the parameters of the surrogate model using both pre-trained and random parameters. Subsequently, we save the models in the intermediate training process to obtain a diverse set of surrogate models. During the HMA stage, we enhance the feature maps of the diversified surrogate models by incorporating beneficial perturbations, thereby further improving the transferability. Experimental results demonstrate that our proposed attack method can effectively enhance the transferability of the crafted adversarial face examples.

cs.CV↗

arXiv2025

STShield: Single-Token Sentinel for Real-Time Jailbreak Detection in Large Language Models

Large Language Models (LLMs) have become increasingly vulnerable to jailbreak attacks that circumvent their safety mechanisms. While existing defense methods either suffer from adaptive attacks or require computationally expensive auxiliary models, we present STShield, a lightweight framework for real-time jailbroken judgement. STShield introduces a novel single-token sentinel mechanism that appends a binary safety indicator to the model's response sequence, leveraging the LLM's own alignment capabilities for detection. Our framework combines supervised fine-tuning on normal prompts with adversarial training using embedding-space perturbations, achieving robust detection while preserving model utility. Extensive experiments demonstrate that STShield successfully defends against various jailbreak attacks, while maintaining the model's performance on legitimate queries. Compared to existing approaches, STShield achieves superior defense performance with minimal computational overhead, making it a practical solution for real-world LLM deployment.

cs.CL↗

arXiv2025

How Far Are We on the Decision-Making of LLMs? Evaluating LLMs' Gaming Ability in Multi-Agent Environments

Decision-making is a complex process requiring diverse abilities, making it an excellent framework for evaluating Large Language Models (LLMs). Researchers have examined LLMs' decision-making through the lens of Game Theory. However, existing evaluation mainly focus on two-player scenarios where an LLM competes against another. Additionally, previous benchmarks suffer from test set leakage due to their static design. We introduce GAMA($γ$)-Bench, a new framework for evaluating LLMs' Gaming Ability in Multi-Agent environments. It includes eight classical game theory scenarios and a dynamic scoring scheme specially designed to quantitatively assess LLMs' performance. $γ$-Bench allows flexible game settings and adapts the scoring system to different game parameters, enabling comprehensive evaluation of robustness, generalizability, and strategies for improvement. Our results indicate that GPT-3.5 demonstrates strong robustness but limited generalizability, which can be enhanced using methods like Chain-of-Thought. We also evaluate 13 LLMs from 6 model families, including GPT-3.5, GPT-4, Gemini, LLaMA-3.1, Mixtral, and Qwen-2. Gemini-1.5-Pro outperforms others, scoring of $69.8$ out of $100$, followed by LLaMA-3.1-70B ($65.9$) and Mixtral-8x22B ($62.4$). Our code and experimental results are publicly available at https://github.com/CUHK-ARISE/GAMABench.

cs.AI↗

arXiv2025

Mitigating Data Scarcity in Time Series Analysis: A Foundation Model with Series-Symbol Data Generation

Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as data scarcity and data imbalance continue to hinder their development. To address this, we consider modeling complex systems through symbolic expressions that serve as semantic descriptors of time series. Building on this concept, we introduce a series-symbol (S2) dual-modulity data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic representations. Leveraging the S2 dataset, we develop SymTime, a pre-trained foundation model for TSA. SymTime demonstrates competitive performance across five major TSA tasks when fine-tuned with downstream task, rivaling foundation models pre-trained on real-world datasets. This approach underscores the potential of dual-modality data generation and pretraining mechanisms in overcoming data scarcity and enhancing task performance.

cs.LG↗

arXiv2025

Tunable Fractional Chern Insulators in Rhombohedral Graphene Superlattices

Fractional Chern insulators (FCIs) showing a transport effect with fractionally quantized Hall plateaus emerging under zero magnetic field, provide a radically new opportunity to engineer topological quantum electronics. By construction of topological flat band with moire engineering, intrinsic FCIs have been observed in twisted MoTe2 system and rhombohedral pentalayer graphene/hBN moire superlattices with anomalous Hall resistivity quantization number C <= 2/3 including the gapless composite Fermi-liquid state with C = 1/2. Here, we experimentally demonstrate a new system of rhombohedral hexalayer graphene (RHG)/hBN moire superlattices, which exhibit both integer and fractional quantum anomalous Hall effects with rich tunability including electric displacement field, perpendicular magnetic field and in-plane magnetic field. By tuning the electrical and magnetic fields at 0 < v < 1, we have observed a quantum phase transition showing a sign reversal of the Hall resistivity at finite magnetic fields. Surprisingly, the FCI state at v = 2/3 survives in the phase transitions, exhibiting a robust quantized Hall resistivity across both phases. Finally we have further demonstrated the indispensable role moire potential plays in the formation of the flat Chern band from a theoretical perspective. Our work has established RHG/hBN moire superlattices as a promising platform for exploring quasi-particles with fractional charge and non-Abelian anyons at zero magnetic field.

cond-mat.mes-hall↗

arXiv2025

BLR-MoE: Boosted Language-Routing Mixture of Experts for Domain-Robust Multilingual E2E ASR

Recently, the Mixture of Expert (MoE) architecture, such as LR-MoE, is often used to alleviate the impact of language confusion on the multilingual ASR (MASR) task. However, it still faces language confusion issues, especially in mismatched domain scenarios. In this paper, we decouple language confusion in LR-MoE into confusion in self-attention and router. To alleviate the language confusion in self-attention, based on LR-MoE, we propose to apply attention-MoE architecture for MASR. In our new architecture, MoE is utilized not only on feed-forward network (FFN) but also on self-attention. In addition, to improve the robustness of the LID-based router on language confusion, we propose expert pruning and router augmentation methods. Combining the above, we get the boosted language-routing MoE (BLR-MoE) architecture. We verify the effectiveness of the proposed BLR-MoE in a 10,000-hour MASR dataset.

cs.CL↗

arXiv2025

QGHNN: A quantum graph Hamiltonian neural network

Representing and learning from graphs is essential for developing effective machine learning models tailored to non-Euclidean data. While Graph Neural Networks (GNNs) strive to address the challenges posed by complex, high-dimensional graph data, Quantum Neural Networks (QNNs) present a compelling alternative due to their potential for quantum parallelism. However, much of the current QNN research tends to overlook the vital connection between quantum state encoding and graph structures, which limits the full exploitation of quantum computational advantages. To address these challenges, this paper introduces a quantum graph Hamiltonian neural network (QGHNN) to enhance graph representation and learning on noisy intermediate-scale quantum computers. Concretely, a quantum graph Hamiltonian learning method (QGHL) is first created by mapping graphs to the Hamiltonian of the topological quantum system. Then, QGHNN based on QGHL is presented, which trains parameters by minimizing the loss function and uses the gradient descent method to learn the graph. Experiments on the PennyLane quantum platform reveal that QGHNN outperforms all assessment metrics, achieving the lowest mean squared error of \textbf{$0.004$} and the maximum cosine similarity of \textbf{$99.8\%$}, which shows that QGHNN not only excels in representing and learning graph information, but it also has high robustness ability. QGHNN can reduce the impact of quantum noise and has significant potential application in future research of quantum knowledge graphs and recommendation systems.

quant-ph↗

arXiv2025

Noise-resistant adaptive Hamiltonian learning

Mitigating and reducing noise influence is crucial for obtaining precise experimental results from noisy intermediate-scale quantum (NISQ) devices. In this work, an adaptive Hamiltonian learning (AHL) model for data analysis and quantum state simulation is proposed to overcome problems such as low efficiency and the noise influence of quantum machine learning algorithms. First, an adaptive parameterized quantum circuit with noise resistant ability is constructed by decomposing the unitary operators that include penalty Hamiltonian in the topological quantum system. Then, a noise-resistant quantum neural network (RQNN) based on AHL is developed, which improves the noise robustness of the quantum neural network by updating iterative parameters. Finally, the experiments on Paddle Quantum demonstrate that RQNN can simulate the mathematical function and get accurate classification results on NISQ devices. Compared with the quantum neural network, RQNN ensures high accuracy with the same non-linear discrete data classification under the impact of amplitude damping noise, with an accuracy of 98.00 $\%$. It provides new possibilities for solving practical issues on NISQ devices and also benefits in the resolution of increasingly complicated problems, which will expand the range of potential applications for quantum machine learning models in the future.

quant-ph↗

arXiv2025

Flat Band and Many-body Gap in Chirally Twisted Triple Bilayer Graphene

We experimentally investigate the band structures of chirally twisted triple bilayer graphene. The new kind of moiré structure, formed by three pieces of helically stacked Bernal bilayer graphene, has flat bands at charge neutral point based on the continuum approximation. We experimentally confirm the existence of flat bands and directly acquire the gap in-between flat bands as well as between the flat bands and dispersive bands from the capacitance measurements. We discover a finite gap even at zero perpendicular electric field, possibly induced by the Coulomb interaction and ferromagnetism. Our quantitative study not only provides solid evidence for the flat-band and interesting physics, but also introduces a quantitative approach to explore phenomena of similar moiré systems.

cond-mat.mes-hall↗

arXiv2024

Diffusion Actor-Critic with Entropy Regulator

Reinforcement learning (RL) has proven highly effective in addressing complex decision-making and control tasks. However, in most traditional RL algorithms, the policy is typically parameterized as a diagonal Gaussian distribution with learned mean and variance, which constrains their capability to acquire complex policies. In response to this problem, we propose an online RL algorithm termed diffusion actor-critic with entropy regulator (DACER). This algorithm conceptualizes the reverse process of the diffusion model as a novel policy function and leverages the capability of the diffusion model to fit multimodal distributions, thereby enhancing the representational capacity of the policy. Since the distribution of the diffusion policy lacks an analytical expression, its entropy cannot be determined analytically. To mitigate this, we propose a method to estimate the entropy of the diffusion policy utilizing Gaussian mixture model. Building on the estimated entropy, we can learn a parameter $α$ that modulates the degree of exploration and exploitation. Parameter $α$ will be employed to adaptively regulate the variance of the added noise, which is applied to the action output by the diffusion model. Experimental trials on MuJoCo benchmarks and a multimodal task demonstrate that the DACER algorithm achieves state-of-the-art (SOTA) performance in most MuJoCo control tasks while exhibiting a stronger representational capacity of the diffusion policy.

cs.LG↗

arXiv2024

MRWeb: An Exploration of Generating Multi-Page Resource-Aware Web Code from UI Designs

Multi-page websites dominate modern web development. However, existing design-to-code methods rely on simplified assumptions, limiting to single-page, self-contained webpages without external resource connection. To address this gap, we introduce the Multi-Page Resource-Aware Webpage (MRWeb) generation task, which transforms UI designs into multi-page, functional web UIs with internal/external navigation, image loading, and backend routing. We propose a novel resource list data structure to track resources, links, and design components. Our study applies existing methods to the MRWeb problem using a newly curated dataset of 500 websites (300 synthetic, 200 real-world). Specifically, we identify the best metric to evaluate the similarity of the web UI, assess the impact of the resource list on MRWeb generation, analyze MLLM limitations, and evaluate the effectiveness of the MRWeb tool in real-world workflows. The results show that resource lists boost navigation functionality from 0% to 66%-80% while facilitating visual similarity. Our proposed metrics and evaluation framework provide new insights into MLLM performance on MRWeb tasks. We release the MRWeb tool, dataset, and evaluation framework to promote further research.

cs.SE↗

arXiv2024

Robust magnetoelectric coupling in altermagnetic-ferroelectric type-III multiferroics

Multiferroic materials, characterized by the coexisting of ferroelectric polarization (breaking spatial inversion symmetry) and magnetism (breaking time-reversal symmetry), with strong magnetoelectric coupling, are highly sought after for advanced technological applications. Novel altermagnets, distinct from conventional magnets, have recently been revealed to exhibit unique spin polarization protected by crystal symmetry, which naturally overcomes the isolation of magnetism from ferroelectrics associated with spatial symmetry. In this study, we propose a novel class of type-III multiferroics, where ferroelectricity and altermagnetism are inherently interlocked by crystal symmetry, setting them apart from conventional multiferroics. Through first-principles calculations, ferroelectric switching is shown to fully invert the spin polarization of altermagnets, equivalent to a 180° reversal of magnetic spin. This strong magnetoelectric coupling is further supported by the magneto-optical Kerr effect, revealing a new class of multiferroics with robust, symmetry-driven magnetoelectric coupling and providing a theoretical foundation for the design of next-generation spintronic devices leveraging altermagnetism.

cond-mat.mtrl-sci↗

arXiv2024

A Spectrum Evaluation Benchmark for Medical Multi-Modal Large Language Models

The significant breakthroughs of Medical Multi-Modal Large Language Models (Med-MLLMs) renovate modern healthcare with robust information synthesis and medical decision support. However, these models are often evaluated on benchmarks that are unsuitable for the Med-MLLMs due to the complexity of real-world diagnostics across diverse specialties. To address this gap, we introduce Asclepius, a novel Med-MLLM benchmark that comprehensively assesses Med-MLLMs in terms of: distinct medical specialties (cardiovascular, gastroenterology, etc.) and different diagnostic capacities (perception, disease analysis, etc.). Grounded in 3 proposed core principles, Asclepius ensures a comprehensive evaluation by encompassing 15 medical specialties, stratifying into 3 main categories and 8 sub-categories of clinical tasks, and exempting overlap with existing VQA dataset. We further provide an in-depth analysis of 6 Med-MLLMs and compare them with 3 human specialists, providing insights into their competencies and limitations in various medical contexts. Our work not only advances the understanding of Med-MLLMs' capabilities but also sets a precedent for future evaluations and the safe deployment of these models in clinical environments.

cs.CL↗

arXiv2024

On the Shortcut Learning in Multilingual Neural Machine Translation

In this study, we revisit the commonly-cited off-target issue in multilingual neural machine translation (MNMT). By carefully designing experiments on different MNMT scenarios and models, we attribute the off-target issue to the overfitting of the shortcuts of (non-centric, centric) language mappings. Specifically, the learned shortcuts biases MNMT to mistakenly translate non-centric languages into the centric language instead of the expected non-centric language for zero-shot translation. Analyses on learning dynamics show that the shortcut learning generally occurs in the later stage of model training, and multilingual pretraining accelerates and aggravates the shortcut learning. Based on these observations, we propose a simple and effective training strategy to eliminate the shortcuts in MNMT models by leveraging the forgetting nature of model training. The only difference from the standard training is that we remove the training instances that may induce the shortcut learning in the later stage of model training. Without introducing any additional data and computational costs, our approach can consistently and significantly improve the zero-shot translation performance by alleviating the shortcut learning for different MNMT models and benchmarks.

cs.CL↗

arXiv2024

Dynamic Textual Prompt For Rehearsal-free Lifelong Person Re-identification

Lifelong person re-identification attempts to recognize people across cameras and integrate new knowledge from continuous data streams. Key challenges involve addressing catastrophic forgetting caused by parameter updating and domain shift, and maintaining performance in seen and unseen domains. Many previous works rely on data memories to retain prior samples. However, the amount of retained data increases linearly with the number of training domains, leading to continually increasing memory consumption. Additionally, these methods may suffer significant performance degradation when data preservation is prohibited due to privacy concerns. To address these limitations, we propose using textual descriptions as guidance to encourage the ReID model to learn cross-domain invariant features without retaining samples. The key insight is that natural language can describe pedestrian instances with an invariant style, suggesting a shared textual space for any pedestrian images. By leveraging this shared textual space as an anchor, we can prompt the ReID model to embed images from various domains into a unified semantic space, thereby alleviating catastrophic forgetting caused by domain shifts. To achieve this, we introduce a task-driven dynamic textual prompt framework in this paper. This model features a dynamic prompt fusion module, which adaptively constructs and fuses two different textual prompts as anchors. This effectively guides the ReID model to embed images into a unified semantic space. Additionally, we design a text-visual feature alignment module to learn a more precise mapping between fine-grained visual and textual features. We also developed a learnable knowledge distillation module that allows our model to dynamically balance retaining existing knowledge with acquiring new knowledge. Extensive experiments demonstrate that our method outperforms SOTAs under various settings.

cs.CV↗

arXiv2024

Understanding and Mitigating the Uncertainty in Zero-Shot Translation

Zero-shot translation is a promising direction for building a comprehensive multilingual neural machine translation~(MNMT) system. However, its quality is still not satisfactory due to off-target issues. In this paper, we aim to understand and alleviate the off-target issues from the perspective of uncertainty in zero-shot translation. By carefully examining the translation output and model confidence, we identify two uncertainties that are responsible for the off-target issues, namely, extrinsic data uncertainty and intrinsic model uncertainty. Based on the observations, we propose two lightweight and complementary approaches to denoise the training data for model training and explicitly penalize the off-target translations by unlikelihood training during model training. Extensive experiments on both balanced and imbalanced datasets show that our approaches significantly improve the performance of zero-shot translation over strong MNMT baselines.

cs.CL↗