Searcharxiv⌕ Search

arXiv subjects

Ziyi Wang

Publications and source records attributed to Ziyi Wang.

At least 145 records · Page 8Linked to original sources
arXiv2025

LVC: A Lightweight Compression Framework for Enhancing VLMs in Long Video Understanding

Long video understanding is a complex task that requires both spatial detail and temporal awareness. While Vision-Language Models (VLMs) obtain frame-level understanding capabilities through multi-frame input, they suffer from information loss due to the sparse sampling strategy. In contrast, Video Large Language Models (Video-LLMs) capture temporal relationships within visual features but are limited by the scarcity of high-quality video-text datasets. To transfer long video understanding capabilities to VLMs with minimal data and computational cost, we propose Lightweight Video Compression (LVC), a novel method featuring the Query-Attention Video Compression mechanism, which effectively tackles the sparse sampling problem in VLMs. By training only the alignment layer with 10k short video-text pairs, LVC significantly enhances the temporal reasoning abilities of VLMs. Extensive experiments show that LVC provides consistent performance improvements across various models, including the InternVL2 series and Phi-3.5-Vision. Notably, the InternVL2-40B-LVC achieves scores of 68.2 and 65.9 on the long video understanding benchmarks MLVU and Video-MME, respectively, with relative improvements of 14.6% and 7.7%. The enhanced models and code will be publicly available soon.

cs.CV↗
arXiv2025

Endo3R: Unified Online Reconstruction from Dynamic Monocular Endoscopic Video

Reconstructing 3D scenes from monocular surgical videos can enhance surgeon's perception and therefore plays a vital role in various computer-assisted surgery tasks. However, achieving scale-consistent reconstruction remains an open challenge due to inherent issues in endoscopic videos, such as dynamic deformations and textureless surfaces. Despite recent advances, current methods either rely on calibration or instrument priors to estimate scale, or employ SfM-like multi-stage pipelines, leading to error accumulation and requiring offline optimization. In this paper, we present Endo3R, a unified 3D foundation model for online scale-consistent reconstruction from monocular surgical video, without any priors or extra optimization. Our model unifies the tasks by predicting globally aligned pointmaps, scale-consistent video depths, and camera parameters without any offline optimization. The core contribution of our method is expanding the capability of the recent pairwise reconstruction model to long-term incremental dynamic reconstruction by an uncertainty-aware dual memory mechanism. The mechanism maintains history tokens of both short-term dynamics and long-term spatial consistency. Notably, to tackle the highly dynamic nature of surgical scenes, we measure the uncertainty of tokens via Sampson distance and filter out tokens with high uncertainty. Regarding the scarcity of endoscopic datasets with ground-truth depth and camera poses, we further devise a self-supervised mechanism with a novel dynamics-aware flow loss. Abundant experiments on SCARED and Hamlyn datasets demonstrate our superior performance in zero-shot surgical video depth prediction and camera pose estimation with online efficiency. Project page: https://wrld.github.io/Endo3R/.

cs.CV↗
arXiv2025

Few-Shot Graph Out-of-Distribution Detection with LLMs

Existing methods for graph out-of-distribution (OOD) detection typically depend on training graph neural network (GNN) classifiers using a substantial amount of labeled in-distribution (ID) data. However, acquiring high-quality labeled nodes in text-attributed graphs (TAGs) is challenging and costly due to their complex textual and structural characteristics. Large language models (LLMs), known for their powerful zero-shot capabilities in textual tasks, show promise but struggle to naturally capture the critical structural information inherent to TAGs, limiting their direct effectiveness. To address these challenges, we propose LLM-GOOD, a general framework that effectively combines the strengths of LLMs and GNNs to enhance data efficiency in graph OOD detection. Specifically, we first leverage LLMs' strong zero-shot capabilities to filter out likely OOD nodes, significantly reducing the human annotation burden. To minimize the usage and cost of the LLM, we employ it only to annotate a small subset of unlabeled nodes. We then train a lightweight GNN filter using these noisy labels, enabling efficient predictions of ID status for all other unlabeled nodes by leveraging both textual and structural information. After obtaining node embeddings from the GNN filter, we can apply informativeness-based methods to select the most valuable nodes for precise human annotation. Finally, we train the target ID classifier using these accurately annotated ID nodes. Extensive experiments on four real-world TAG datasets demonstrate that LLM-GOOD significantly reduces human annotation costs and outperforms state-of-the-art baselines in terms of both ID classification accuracy and OOD detection performance.

cs.LG↗
arXiv2025

Cauchy-Schwarz Regularizers

We introduce a novel class of regularization functions, called Cauchy-Schwarz (CS) regularizers, which can be designed to induce a wide range of properties in solution vectors of optimization problems. To demonstrate the versatility of CS regularizers, we derive regularization functions that promote discrete-valued vectors, eigenvectors of a given matrix, and orthogonal matrices. The resulting CS regularizers are simple, differentiable, and can be free of spurious stationary points, making them suitable for gradient-based solvers and large-scale optimization problems. In addition, CS regularizers automatically adapt to the appropriate scale, which is, for example, beneficial when discretizing the weights of neural networks. To demonstrate the efficacy of CS regularizers, we provide results for solving underdetermined systems of linear equations and weight quantization in neural networks. Furthermore, we discuss specializations, variations, and generalizations, which lead to an even broader class of new and possibly more powerful regularizers.

math.OC↗
arXiv2025

Safe-VAR: Safe Visual Autoregressive Model for Text-to-Image Generative Watermarking

With the success of autoregressive learning in large language models, it has become a dominant approach for text-to-image generation, offering high efficiency and visual quality. However, invisible watermarking for visual autoregressive (VAR) models remains underexplored, despite its importance in misuse prevention. Existing watermarking methods, designed for diffusion models, often struggle to adapt to the sequential nature of VAR models. To bridge this gap, we propose Safe-VAR, the first watermarking framework specifically designed for autoregressive text-to-image generation. Our study reveals that the timing of watermark injection significantly impacts generation quality, and watermarks of different complexities exhibit varying optimal injection times. Motivated by this observation, we propose an Adaptive Scale Interaction Module, which dynamically determines the optimal watermark embedding strategy based on the watermark information and the visual characteristics of the generated image. This ensures watermark robustness while minimizing its impact on image quality. Furthermore, we introduce a Cross-Scale Fusion mechanism, which integrates mixture of both heads and experts to effectively fuse multi-resolution features and handle complex interactions between image content and watermark patterns. Experimental results demonstrate that Safe-VAR achieves state-of-the-art performance, significantly surpassing existing counterparts regarding image quality, watermarking fidelity, and robustness against perturbations. Moreover, our method exhibits strong generalization to an out-of-domain watermark dataset QR Codes.

cs.MM↗
arXiv2025

SemGeoMo: Dynamic Contextual Human Motion Generation with Semantic and Geometric Guidance

Generating reasonable and high-quality human interactive motions in a given dynamic environment is crucial for understanding, modeling, transferring, and applying human behaviors to both virtual and physical robots. In this paper, we introduce an effective method, SemGeoMo, for dynamic contextual human motion generation, which fully leverages the text-affordance-joint multi-level semantic and geometric guidance in the generation process, improving the semantic rationality and geometric correctness of generative motions. Our method achieves state-of-the-art performance on three datasets and demonstrates superior generalization capability for diverse interaction scenarios. The project page and code can be found at https://4dvlab.github.io/project_page/semgeomo/.

cs.CV↗
arXiv2025

Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning

Reasoning abilities, especially those for solving complex math problems, are crucial components of general intelligence. Recent advances by proprietary companies, such as o-series models of OpenAI, have made remarkable progress on reasoning tasks. However, the complete technical details remain unrevealed, and the techniques that are believed certainly to be adopted are only reinforcement learning (RL) and the long chain of thoughts. This paper proposes a new RL framework, termed OREAL, to pursue the performance limit that can be achieved through \textbf{O}utcome \textbf{RE}w\textbf{A}rd-based reinforcement \textbf{L}earning for mathematical reasoning tasks, where only binary outcome rewards are easily accessible. We theoretically prove that behavior cloning on positive trajectories from best-of-N (BoN) sampling is sufficient to learn the KL-regularized optimal policy in binary feedback environments. This formulation further implies that the rewards of negative samples should be reshaped to ensure the gradient consistency between positive and negative samples. To alleviate the long-existing difficulties brought by sparse rewards in RL, which are even exacerbated by the partial correctness of the long chain of thought for reasoning tasks, we further apply a token-level reward model to sample important tokens in reasoning trajectories for learning. With OREAL, for the first time, a 7B model can obtain 94.0 pass@1 accuracy on MATH-500 through RL, being on par with 32B models. OREAL-32B also surpasses previous 32B models trained by distillation with 95.0 pass@1 accuracy on MATH-500. Our investigation also indicates the importance of initial policy models and training queries for RL. Code, models, and data will be released to benefit future research\footnote{https://github.com/InternLM/OREAL}.

cs.CL↗
arXiv2025

SelectIT: Selective Instruction Tuning for LLMs via Uncertainty-Aware Self-Reflection

Instruction tuning (IT) is crucial to tailoring large language models (LLMs) towards human-centric interactions. Recent advancements have shown that the careful selection of a small, high-quality subset of IT data can significantly enhance the performance of LLMs. Despite this, common approaches often rely on additional models or data, which increases costs and limits widespread adoption. In this work, we propose a novel approach, termed SelectIT, that capitalizes on the foundational capabilities of the LLM itself. Specifically, we exploit the intrinsic uncertainty present in LLMs to more effectively select high-quality IT data, without the need for extra resources. Furthermore, we introduce a curated IT dataset, the Selective Alpaca, created by applying SelectIT to the Alpaca-GPT4 dataset. Empirical results demonstrate that IT using Selective Alpaca leads to substantial model ability enhancement. The robustness of SelectIT has also been corroborated in various foundation models and domain-specific tasks. Our findings suggest that longer and more computationally intensive IT data may serve as superior sources of IT, offering valuable insights for future research in this area. Data, code, and scripts are freely available at https://github.com/Blue-Raincoat/SelectIT.

cs.CL↗
arXiv2024

Demystify Mamba in Vision: A Linear Attention Perspective

Mamba is an effective state space model with linear computation complexity. It has recently shown impressive efficiency in dealing with high-resolution inputs across various vision tasks. In this paper, we reveal that the powerful Mamba model shares surprising similarities with linear attention Transformer, which typically underperform conventional Transformer in practice. By exploring the similarities and disparities between the effective Mamba and subpar linear attention Transformer, we provide comprehensive analyses to demystify the key factors behind Mamba's success. Specifically, we reformulate the selective state space model and linear attention within a unified formulation, rephrasing Mamba as a variant of linear attention Transformer with six major distinctions: input gate, forget gate, shortcut, no attention normalization, single-head, and modified block design. For each design, we meticulously analyze its pros and cons, and empirically evaluate its impact on model performance in vision tasks. Interestingly, the results highlight the forget gate and block design as the core contributors to Mamba's success, while the other four designs are less crucial. Based on these findings, we propose a Mamba-Inspired Linear Attention (MILA) model by incorporating the merits of these two key designs into linear attention. The resulting model outperforms various vision Mamba models in both image classification and high-resolution dense prediction tasks, while enjoying parallelizable computation and fast inference speed. Code is available at https://github.com/LeapLabTHU/MLLA.

cs.CV↗
arXiv2024

Global and Local Attention-Based Transformer for Hyperspectral Image Change Detection

Recently Transformer-based hyperspectral image (HSI) change detection methods have shown remarkable performance. Nevertheless, existing attention mechanisms in Transformers have limitations in local feature representation. To address this issue, we propose Global and Local Attention-based Transformer (GLAFormer), which incorporates a global and local attention module (GLAM) to combine high-frequency and low-frequency signals. Furthermore, we introduce a cross-gating mechanism, called cross-gated feed-forward network (CGFN), to emphasize salient features and suppress noise interference. Specifically, the GLAM splits attention heads into global and local attention components to capture comprehensive spatial-spectral features. The global attention component employs global attention on downsampled feature maps to capture low-frequency information, while the local attention component focuses on high-frequency details using non-overlapping window-based local attention. The CGFN enhances the feature representation via convolutions and cross-gating mechanism in parallel paths. The proposed GLAFormer is evaluated on three HSI datasets. The results demonstrate its superiority over state-of-the-art HSI change detection methods. The source code of GLAFormer is available at \url{https://github.com/summitgao/GLAFormer}.

eess.IV↗
arXiv2024

XMask3D: Cross-modal Mask Reasoning for Open Vocabulary 3D Semantic Segmentation

Existing methodologies in open vocabulary 3D semantic segmentation primarily concentrate on establishing a unified feature space encompassing 3D, 2D, and textual modalities. Nevertheless, traditional techniques such as global feature alignment or vision-language model distillation tend to impose only approximate correspondence, struggling notably with delineating fine-grained segmentation boundaries. To address this gap, we propose a more meticulous mask-level alignment between 3D features and the 2D-text embedding space through a cross-modal mask reasoning framework, XMask3D. In our approach, we developed a mask generator based on the denoising UNet from a pre-trained diffusion model, leveraging its capability for precise textual control over dense pixel representations and enhancing the open-world adaptability of the generated masks. We further integrate 3D global features as implicit conditions into the pre-trained 2D denoising UNet, enabling the generation of segmentation masks with additional 3D geometry awareness. Subsequently, the generated 2D masks are employed to align mask-level 3D representations with the vision-language feature space, thereby augmenting the open vocabulary capability of 3D geometry embeddings. Finally, we fuse complementary 2D and 3D mask features, resulting in competitive performance across multiple benchmarks for 3D open vocabulary semantic segmentation. Code is available at https://github.com/wangzy22/XMask3D.

cs.CV↗
arXiv2024

Theoretical and Experimental Study on Heat Transfer Characteristics of Water Heat Pipe

Heat pipe is an efficient heat transfer element based on two-phase natural circulation, which has advantages of simple structure, strong heat transfer ability, and good isothermal performance. Heat pipes are widely used in heat transfer and other fields, and especially have important applications in nuclear engineering. One of its most important characteristics is the heat transfer limit. In this work, heat transfer limits are first reviewed, and the detailed calculation equations are presented. Then, a Matlab code to calculate heat transfer limits as well as the thermal conductance are provided. Second, an experimental setup for testing the heat transfer characteristics of heat pipes was developed, which could be used to measure the thermal conductance as well as the heat transfer limits. The calculated results show that, for water heat pipes, the capillary limit mainly affects heat transfer in low temperature conditions, while in high temperature conditions, boiling limit dominates. In addition, the experiment results show that the thermal conductance of the measured heat pipe is 7267 W/(m^2*K), which agrees with the calculation result.

physics.app-ph↗
arXiv2024

Emotion-Aware Interaction Design in Intelligent User Interface Using Multi-Modal Deep Learning

In an era where user interaction with technology is ubiquitous, the importance of user interface (UI) design cannot be overstated. A well-designed UI not only enhances usability but also fosters more natural, intuitive, and emotionally engaging experiences, making technology more accessible and impactful in everyday life. This research addresses this growing need by introducing an advanced emotion recognition system to significantly improve the emotional responsiveness of UI. By integrating facial expressions, speech, and textual data through a multi-branch Transformer model, the system interprets complex emotional cues in real-time, enabling UIs to interact more empathetically and effectively with users. Using the public MELD dataset for validation, our model demonstrates substantial improvements in emotion recognition accuracy and F1 scores, outperforming traditional methods. These findings underscore the critical role that sophisticated emotion recognition plays in the evolution of UIs, making technology more attuned to user needs and emotions. This study highlights how enhanced emotional intelligence in UIs is not only about technical innovation but also about fostering deeper, more meaningful connections between users and the digital world, ultimately shaping how people interact with technology in their daily lives.

cs.HC↗
arXiv2024

Zeolitic Imidazolate Framework-8 offers an anti-inflammatory and antifungal method in the treatment of Aspergillus fungus keratitis in vitro and in vivo

Background: Fungal keratitis is a serious blinding eye disease. Traditional drugs used to treat fungal keratitis commonly have the disadvantages of low bioavailability, poor dispersion, and limited permeability. Purpose: To develop a new method for the treatment of fungal keratitis with improved bioavailability, dispersion, and permeability. Purpose: To develop a new method for the treatment of fungal keratitis with improved bioavailability, dispersion, and permeability. Methods: Zeolitic Imidazolate Framework-8 (ZIF-8) was formed by zinc ions and 2-methylimidazole linked by coordination bonds and characterized by Scanning electron microscopy (SEM), X-ray diffraction (XRD), and Zeta potential. The safety of ZIF-8 on HCECs and RAW 264.7 cells was detected by Cell Counting Kit-8 (CCK-8). The anti-inflammatory effects of ZIF-8 on RAW 246.7 cells were evaluated by Quantitative Real-Time PCR Experiments (qPCR) and Enzyme-linked immunosorbent assay (ELISA). Clinical score, Colony-Forming Units (CFU). In vivo, treatment with ZIF-8 reduced corneal fungal load and mitigated neutrophil infiltration in fungal keratitis, which effectively reduced the severity of keratitis in mice and alleviated the infiltration of inflammatory factors in the mouse cornea. In addition, ZIF-8 reduces the inflammatory response by downregulating the expression of pro-inflammatory cytokines TNF-α, IL-6, and IL-1\b{eta} after Aspergillus fumigatus infection in vivo and in vitro. Conclusion: ZIF-8 has a significant anti-inflammatory and antifungal effect, which provides a new solution for the treatment of fungal keratitis.

q-bio.TO↗
arXiv2024

ZIF-90 treats fungal keratitis by promoting macrophage apoptosis and inhibiting inflammatory response

Fungal keratitis is a severe vision-threatening corneal infection with a prognosis influenced by fungal virulence and the host's immune defense mechanisms. The immune system, through its regulation of the inflammatory response, ensures cells and tissues can effectively activate defense mechanisms in response to infection and injury. However, there is still a lack of effective drugs that attenuate fungal virulence while relieving the inflammatory response caused by fungal keratitis. Therefore, finding effective treatments to solve these problems is particularly important. We synthesized ZIF-90 by water-based synthesis and characterized by SEM, XRD etc. In vitro experiments included CCK-8 and ELISA. These evaluations verified the disruptive effects of ZIF-90 on Aspergillus. fumigatus spore adhesion, morphology, cell membrane, and the effect of ZIF-90 on apoptosis. In addition, to investigate whether the metal-ligand zinc and the organic ligand imidazole act as essential factors in ZIF-90, we investigated the in vitro antimicrobial and anti-inflammatory effects of ZIF-8, ZIF-67, and MOF-74 (Zn) by MIC and ELISA experiments. ZIF-90 has therapeutic effects on fungal keratitis, which could break the protective organelles of Aspergillus. fumigatus, such as the cell wall. In addition, ZIF-90 can avoid excessive inflammatory response by promoting apoptosis of inflammatory cells. The results demonstrated that both zinc ions and imidazole possessed antimicrobial and anti-inflammatory effects. In addition, ZIF-90 exhibited better biocompatibility compared to ZIF-8, ZIF-67, and MOF-74 (Zn). ZIF-90 has anti-inflammatory and antifungal effects and preferable biocompatibility, and has great potential for the treatment of fungal keratitis.

q-bio.SC↗
arXiv2024

Efficient and Aesthetic UI Design with a Deep Learning-Based Interface Generation Tree Algorithm

This paper presents a novel method for user interface (UI) generation based on the Transformer architecture, addressing the increasing demand for efficient and aesthetically pleasing UI designs in software development. Traditional UI design relies heavily on designers' expertise, which can be time-consuming and costly. Leveraging the capabilities of Transformers, particularly their ability to capture complex design patterns and long-range dependencies, we propose a Transformer-based interface generation tree algorithm. This method constructs a hierarchical representation of UI components as nodes in a tree structure, utilizing pre-trained Transformer models for encoding and decoding. We define a markup language to describe UI components and their properties and use a rich dataset of real-world web and mobile application interfaces for training. The experimental results demonstrate that our approach not only significantly enhances design quality and efficiency but also outperforms traditional models in user satisfaction and aesthetic appeal. We also provide a comparative analysis with existing models, illustrating the advantages of our method in terms of accuracy, user ratings, and design similarity. Overall, our study underscores the potential of the Transformer based approach to revolutionize the UI design process, making it accessible for non-professionals while maintaining high standards of quality.

cs.HC↗