检索结果-内蒙古大学图书馆

Impacts of Vertical Resolution and Diurnally Varying Background Error Covariance on FY-4A/AGRI data Assimilation over Land

引用

Monthly Weather Review 2025年第5期153卷 763-790页

作者： Li, Xin Zou, Xiaolei Xu, Xu Liu, Weiguang Nanjing Innovation Institute for Atmospheric Sciences Chinese Academy of Meteorological Sciences Jiangsu Meteorological Service Nanjing China Key Laboratory of Transportation Meteorology of CMA Jiangsu Key Laboratory of Severe Storm Disaster Risk Nanjing China Center of Data Assimilation for Research and Application Nanjing University of Information Science and Technology Nanjing China

This study aims at improving the assimilation of Fengyun-4A (FY-4A)/Advanced Geostationary Radiation Imager (AGRI) clear-sky surface-sensitive brightness temperature over land using the Weather Research and Forecasting and the Gridpoint Statistical Interpolation three-dimensional variational data assimilation for summer convective cases over eastern China. We investigated the influences of the planetary boundary layer on the assimilation of FY-4A/AGRI surface-sensitive channels, focusing on model low-level vertical resolution and diurnal variation of background error covariance. The vertical levels are increased from 48 to 91 by adding more near-surface levels. First, stronger low-level vertical background error correlations of humidity and temperature are found during daytime than during nighttime. A total of five numerical experiments are then conducted to examine the impacts of FY-4A/AGRI assimilation with two vertical resolutions. The experiments WV48 and WV91 assimilate only the water vapor channels, and ALL48, ALL91, and ALL91M assimilate the water vapor and surface-sensitive channels. The diurnally varying background error covariance, which is estimated at 3-h intervals, is incorporated into ALL91M. The 6-h cycling data assimilation at 1-h interval prior to convection initiation is conducted for each of the seven cases. Overall, a comparison between WV48/WV91 and ALL48/ALL91M revealed a consistent positive impact of AGRI brightness temperature assimilation over land on convective precipitation forecasts. The added values of assimilating AGRI surface-sensitive channels are more pronounced by using the finer vertical resolution due to low-level humidity enhancements that are related to the rise of the planetary boundary layer. With the diurnally varying background error covariance, ALL91M outperformed ALL91 because of improved low-level analyzed structures that are associated with proper low-level vertical error correlations of humidity and thermal variables. ©

关键词： Snow

来源：评论

学校读者我要写书评

暂无评论

Improving Transferable Targeted Adversarial Attack via Normalized Logit Calibration and Truncated Feature Mixing

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 4595-4609页

作者： Weng, Juanjuan Luo, Zhiming Li, Shaozi Jinan University College of Information Science and Technology Guangzhou510632 China Xiamen University Department of Artificial Intelligence Xiamen361005 China Wuyi University Fujian Key Laboratory of Big Data Application and Intellectualization for Tea Industry Wuyishan354300 China

This paper aims to enhance the transferability of adversarial samples in targeted attacks, where attack success rates remain comparatively low. To achieve this objective, we propose two distinct techniques for improving the targeted transferability from the loss and feature aspects. First, in previous approaches, logit calibrations used in targeted attacks primarily focus on the logit margin between the targeted class and the untargeted classes among samples, neglecting the standard deviation of the logit. In contrast, we introduce a new normalized logit calibration method that jointly considers the logit margin and the standard deviation of logits. This approach effectively calibrates the logits, enhancing the targeted transferability. Second, previous studies have demonstrated that mixing the features of clean samples during optimization can significantly increase transferability. Building upon this, we further investigate a truncated feature mixing method to reduce the impact of the source training model, resulting in additional improvements. The truncated feature is determined by removing the Rank-1 feature associated with the largest singular value decomposed from the high-level convolutional layers of the clean sample. Extensive experiments conducted on the ImageNet-Compatible, CIFAR-10 and ImageNet-1k datasets demonstrate the individual and mutual benefits of our proposed two components, which outperform the state-of-the-art methods by a large margin in black-box targeted attacks. © 2005-2012 IEEE.

关键词： Calibration Closed box Training Standards Translation Perturbation methods Optimization Predictive models Feature extraction Industries

来源：评论

学校读者我要写书评

暂无评论

High-strength Al2O3-Ti(C,N) composites fabricated by sinter forging

引用

Ceramics International 2025年第14期51卷 18956-18965页

作者： Zhang, Mengwen Meng, Fanyu Jiang, Binyang Li, Jinsheng Fan, Lei Yang, Shoulei Henan Key Laboratory of Aeronautical Materials and Application Technology Zhengzhou University of Aeronautics Henan Zhengzhou450046 China School of Materials Science and Engineering Zhengzhou University of Aeronautics Henan Zhengzhou450046 China Key Laboratory of Dielectric and Electrolyte Functional Materials Hebei Province School of Resources and Materials Northeastern University at Qinhuangdao Qinhuangdao066004 China

To further enhance the performance of Al2O3-Ti(C,N) composites, this study employed novel sinter forging and dynamic sinter forging techniques, comparing them with conventional hot pressing and oscillatory pressure sintering. The results revealed that the composites produced by all four processes exhibited high density and no phase transition or formation of new phases. However, the composites prepared by sinter forging and dynamic sinter forging exhibited significantly improved strength and hardness. The hardness of the composites produced by sinter forging increased to 23.39 GPa, while the strength reached 1142 MPa, representing a 61 % improvement compared to those sintered by hot pressing. This improvement can be attributed to the significant plastic deformation induced by the sinter-forging process, which results in a higher dislocation density at the grain boundaries and within certain grains, enhanced coordinated deformation between grains, and stronger grain boundary bonding. Furthermore, the average grain size was smaller. Therefore, sinter forging and dynamic sinter forging offer substantial advantages in the preparation of advanced ceramics. © 2025 Elsevier Ltd and Techna Group S.r.l.

关键词： Brinell Hardness

来源：评论

学校读者我要写书评

暂无评论

StoryLLaVA: Enhancing Visual Storytelling with Multi-Modal Large Language Models 31

StoryLLaVA: Enhancing Visual Storytelling with Multi-Modal L...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Yang, Li Xiao, Zhiding Huang, Wenxin Zhong, Xian Hubei Key Laboratory of Big Data Intelligent Analysis and Application School of Computer Science and Information Engineering Hubei University China Hubei Key Laboratory of Transportation Internet of Things School of Computer Science and Artificial Intelligence Wuhan University of Technology China

ISBN: (纸本)9798891761964

The rapid development of multi-modal large language models (MLLMs) has positioned visual storytelling as a crucial area in content creation. However, existing models often struggle to maintain temporal, spatial, and narrative coherence across image sequences and frequently lack the depth and engagement of human-authored stories. To address these challenges, we propose Story with Large Language and Vision Assistant (StoryLLaVA), a novel framework for enhancing visual storytelling. Our approach introduces a Topic-Driven Narrative Optimizer (TDNO) that improves both the training data and MLLM models by integrating image descriptions, topic generation, and GPT-4-based refinements. Furthermore, we employ a preference-based ranked story sampling method that aligns model outputs with human storytelling preferences through positive-negative pairing. These two phases of the framework differ in their training methods: the former uses supervised fine-tuning, while the latter incorporates reinforcement learning with positive and negative sample pairs. Experimental results demonstrate that StoryLLaVA outperforms current models in visual relevance, coherence, and fluency, with LLM-based evaluations confirming the generation of richer and more engaging narratives. The enhanced dataset and model are available at https://***/XxxZzD/StoryLLaVA. © 2025 Association for Computational Linguistics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Dual Test-Time Training for Out-of-Distribution Recommender System

引用

IEEE Transactions on Knowledge and data Engineering 2025年第6期37卷 3312-3326页

作者： Yang, Xihong Wang, Yiqi Chen, Jin Fan, Wenqi Zhao, Xiangyu Zhu, En Liu, Xinwang Lian, Defu National University of Defense Technology College of Computer Science and Technology Changsha410073 China The Hong Kong University of Science and Technology School of Business and Management Hong Kong Hong Kong City University of Hong Kong Department of Data Science Hong Kong University of Science and Technology of China Anhui Province Key Laboratory of Big Data Analysis and Application School of Computer Science and Technology Hefei Anhui230000 China

Deep learning has been widely applied in recommender systems, which has recently achieved revolutionary progress. However, most existing learning-based methods assume that the user and item distributions remain unchanged between the training phase and the test phase. However, the distribution of user and item features can naturally shift in real-world scenarios, potentially resulting in a substantial decrease in recommendation performance. This phenomenon can be formulated as an Out-Of-Distribution (OOD) recommendation problem. To address this challenge, we propose a novel Dual Test-Time-Training framework for OOD Recommendation, termed DT3OR. In DT3OR, we incorporate a model adaptation mechanism during the test-time phase to carefully update the recommendation model, allowing the model to adapt specially to the shifting user and item features. To be specific, we propose a self-distillation task and a contrastive task to assist the model learning both the user's invariant interest preferences and the variant user/item characteristics during the test-time phase, thus facilitating a smooth adaptation to the shifting features. Furthermore, we provide theoretical analysis to support the rationale behind our dual test-time training framework. To the best of our knowledge, this paper is the first work to address OOD recommendation via a test-time-training strategy. We conduct experiments on five datasets with various backbones. Comprehensive experimental results have demonstrated the effectiveness of DT3OR compared to other state-of-the-art baselines. © 2025 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Robust adaptive distributed optimization for heterogeneous unknown second-order nonlinear multiagent systems

引用

science China(Information sciences) 2025年第4期68卷 391-392页

作者： Ping GONG Qing-Guo WANG School of Mathematics and Statistics Guangdong University of Foreign Studies Institute of Artificial Intelligence and Future Networks Beijing Normal University Guangdong Provincial/Zhuhai Key Laboratory of Interdisciplinary Research and Application for Data Science Beijing Normal University-Hong Kong Baptist University United International College

The distributed optimization problem has drawn significant attention recently since it has a variety of applications in the economic dispatch of smart grids, regression of distributed data, optimal coordination of mobile robot systems(MRSs),and so on. Different from the traditional leaderless consensus and consensus tracking problems in multiagent systems(MASs), the typical distributed optimization problems are not only required to achieve consensus but also to cooperatively minimize a global cost function [1], where the global cost function is the sum of local cost functions and each local cost function is only known to the agent itself.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lightweight Image Encryption Algorithm Using 4D-NDS: Compound Dynamic Diffusion and Single-round Efficiency

引用

IEEE Access 2025年 13卷 74656-74666页

作者： Liao, Yunlong Lin, Yiting Li, Qiutong Xing, Zheng Yuan, Xiaochen Macao Polytechnic University Faculty of Applied Sciences Macao999078 China Beijing Normal-Hong Kong Baptist University Guangdong Provincial Key Laboratory of Interdisciplinary Research and Application for Data Science Zhuhai519087 China Hong Kong Baptist University Department of Computer Science Hong Kong Hong Kong

With the advancement of the Internet of Things (IoT) and communication technologies, multimedia information has become a significant component of communication. Image data, as a carrier with high information density and large transmission volume, is increasingly at risk of information leakage. To address the issue of digital private image leakage, this paper proposes a reliable image security protection scheme based on non-degenerate hyper-chaotic systems for dynamic permutation and single-round combined diffusion. By leveraging the high dynamic properties of non-degenerate chaos, the scheme effectively resolves the insufficient sensitivity of pseudo-random sequences to initial values. Additionally, the single-round combined double-diffusion system significantly enhances the encryption efficiency of the algorithm. Experimental and theoretical verification demonstrate that the algorithm exhibits high sensitivity and effective robustness, capable of resisting common attacks and providing a reliable security solution for communication and IoT. © 2013 IEEE.

关键词： Information leakage

来源：评论

学校读者我要写书评

暂无评论

Byzantine-robust distributed support vector machine

引用

science China Mathematics 2025年第3期68卷 707-728页

作者： Xiaozhou Wang Weidong Liu Xiaojun Mao School of Statistics East China Normal UniversityShanghai 200062China Key Laboratory of Advanced Theory and Application in Statistics and Data Science(Ministry of Education) East China Normal UniversityShanghai 200062China School of Mathematical Sciences Shanghai Jiao Tong UniversityShanghai 200240China Key Laboratory of Artificial Intelligence(Ministry of Education) Shanghai Jiao Tong UniversityShanghai 200240China Key Laboratory of Scientific and Engineering Computing(Ministry of Education) Shanghai Jiao Tong UniversityShanghai 200240China

The development of information technology brings diversification of data sources and large-scale data sets and calls for the exploration of distributed learning algorithms. In distributed systems, some local machines may behave abnormally and send arbitrary information to the central machine(known as Byzantine failures), which can invalidate the distributed algorithms based on the assumption of faultless systems. This paper studies Byzantine-robust distributed algorithms for support vector machines(SVMs) in the context of binary classification. Despite a vast literature on Byzantine problems, much less is known about the theoretical properties of Byzantine-robust SVMs due to their unique challenges. In this paper, we propose two distributed gradient descent algorithms for SVMs. The median and trimmed mean operations in aggregation can effectively defend against Byzantine failures. Theoretically, we show the convergence of the proposed estimators and provide the statistical error rates. After a certain number of iterations, our estimators achieve near-optimal rates. Simulation studies and real data analysis are conducted to demonstrate the performance of the proposed Byzantine-robust distributed algorithms.

关键词： Byzantine robustness convergence distributed learning support vector machine

来源：评论

学校读者我要写书评

暂无评论

GBC: Gaussian-Based Colorization and Super-Resolution for 3D Reconstruction 19

GBC: Gaussian-Based Colorization and Super-Resolution for 3D...

引用

19th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and its applications in Industry, VRCAI 2024

作者： Liao, Jiecheng He, Shi Yuan, Yichen Zhang, Hui Department of Computer Science and Technology BNU-HKBU United International College ZhuHai China Guangdong Provincial Key Laboratory of Interdisciplinary Research and Application for Data Science BNU-HKBU United International College Zhuhai China

ISBN: (纸本)9798400713484

In this paper, we introduce GBC, an advanced framework for transforming low-resolution, monochrome video sequences into high-resolution, colorized, and geometrically accurate 3D models. GBC combines Bidirectional Optical Flow Super-Resolution (BOF-SR) for temporal super-resolution with a novel colorization approach, Temporal Optical Flow-based Colorization (TOF-CO), designed to enhance frame-to-frame temporal consistency. The integration of these modules with COLMAP-based 3D Gaussian splatting further extends GBC's capability to reconstruct high-fidelity 3D scenes. Additionally, we created a custom dataset tailored to the challenges of low-light, low-quality historical footage, enabling robust evaluation alongside public datasets. This solution offers a new approach to video restoration, advancing temporal coherence, color accuracy, and 3D scene fidelity. The source code is publicly available at https://***/ffftuanxxx/GBC. © 2024 ACM.

关键词： 3D reconstruction

来源：评论

学校读者我要写书评

暂无评论

High-density low-angle grain boundaries enable exceptional combination of strength, ductility and toughness in 304 stainless steel

引用

Journal of Materials science and Technology 2025年 238卷 191-198页

作者： Fan, Kuanyuan Liu, Baoxi Yin, Fuxing He, Jiazhen He, Binbin Luo, Zhichao Research Institute for Energy Equipment Materials Tianjin key laboratory of materials laminating fabrication and interfacial controlling technology School of Materials Science and Engineering Hebei University of Technology Tianjin 300132 China Institute of New Materials Guangdong Academy of Science Guangdong Provincial Key Laboratory of Metal Toughening Technology and Application Guangzhou510651 China Department of Mechanical and Energy Engineering Southern University of Science and Technology Shenzhen518055 China

Practical methodologies for enhancing the strength, ductility, and toughness of structural metals remain a significant challenge. In this study, we propose a low-angle grain boundary (LAGB) engineering protocol that utilizes a continuous dynamic recrystallization (CDRX) mechanism during warm deformation, specifically applied to 304 austenitic stainless steel (ASS). We demonstrate for the first time that the high density of LAGBs contributes approximately 212 MPa to the yield strength of 304 ASS. Furthermore, the introduction of LAGBs enhances the stability of austenite due to its high mobility and interface stability, which delays the transformation-induced plasticity (TRIP) effect, resulting in increased tensile elongation and toughness. © 2025

关键词： Austenitic stainless steel

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：