检索结果-内蒙古大学图书馆

IEEE Conference on Vehicular Technology (VTC)

作者： Zhihao Chen Xinhua Zheng Xiang Chen School of Electronics and Information Technology Sun Yat-sen University Guangzhou China Guangdong Provincial Key Laboratory of Big Data Computing The Chinese University of Hong Kong Shenzhen China

In the OTFS channel estimation scheme, the spectral efficiency loss can be mitigated by a superimposed pilot (SP) scheme, yet it compromises the peak-to-average power ratio (PAPR) performance. This paper introduces a delay-wise superimposed pilot placement scheme, termed the SP-DW scheme, based on Zadoff-Chu (ZC) sequences. This scheme is designed to enhance the PAPR performance of the system. Specifically, the ZC sequences can maintain good mutual incoherence between the column vectors of the dictionary matrix in compressed sensing-based channel estimation, thereby averting estimation failure. Additionally, a two-step channel estimator and iterative scheme based on the multipath matching pursuit (MMP) algorithm are developed to search for fractional delays and Doppler shifts. The MMP algorithm optimally combines multiple paths in the fractional dictionary matrix to identify sparse candidates, subsequently improving the accuracy of channel estimation. Simulation results indicate that the proposed scheme outperforms the existing SP scheme in terms of PAPR, channel estimation, and bit error rate (BER) performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Some Support Vector Regression Machines with Given Empirical Risks Partly

引用

Journal of Computers (Taiwan) 2022年第5期33卷 61-72页

作者： Luo, Lin-Kai Xu, Chao-Jie Ye, Ling-Jun Peng, Hong Department of Automation Xiamen University Xiamen China Xiamen Key Laboratory of Intelligent Analysis and Decision of Big Data Xiamen China

There are often some prior requirements about empirical risk in regression problems. To meet these requirements, this paper firstly proposes two novel support vector regression machine models in which part of empirical risks are given. One is a support vector regression machine in which partial empirical risks are given (PSVR), and the other is a model in which unilateral partial empirical risks are given (UPSVR). For the samples with given empirical risk levels, PSVR meets the requirements by some inequality constraints about empirical risk levels, while for the other samples without empirical risk requirement, PSVR uses the same strategy as the tradition support vector regression (SVR) to meet the requirement of empirical risk. UPSVR is similar to PSVR, except that the inequality constrains of empirical risks are unilateral. Secondly, the dual problems and the solving methods of PSVR and UPSVR are given. Finally, the effectiveness and superiority of PSVR and UPSVR are verified by the experiments on four artificial datasets. Both PSVR and UPSVR achieve better regression performance than the traditional models respectively. At the same time, PSVR is less sensitive to the trade-off coefficient C between empirical risk and confident risk compared with SVR. Thus, PSVR can select parameter C faster and more conveniently. PSVR and UPSVR are the extensions of the traditional models. When the set of samples with given empirical risks is empty, they degenerate into the traditional models. PSVR and UPSVR are suitable for the scene with prior requirements of empirical risk. © 2022 Computer Society of the Republic of China. All rights reserved.

关键词： Economic and social effects

来源：评论

学校读者我要写书评

暂无评论

Hyper-Parameter Optimization and Proxy Re-encryption for Federated Learning 6th

Hyper-Parameter Optimization and Proxy Re-encryption for F...

引用

6th International Conference on Blockchain and Trustworthy Systems, BlockSys 2024

作者： Du, Wanrong Wen, Bin Ma, Mengshuai Liu, Wenlong Wei, Xiaoxun Key Laboratory of Data Science and Smart Education Ministry of Education Hainan Normal University Haikou571158 China School of Information Science and Technology Hainan Normal University Haikou571158 China Cloud Computing and Big Data Research Center Hainan Normal University Haikou571158 China

ISBN: (纸本)9789819614103

Federated learning is a distributed machine learning paradigm that allows multiple participants to collaborate on training a shared model without revealing their own data. However, federated learning faces challenges such as data heterogeneity, communication overhead, security attacks, etc., which result in low efficiency of data training. To address these issues, this paper proposes a strategy to enhance the training efficiency of federated learning data through hyperparameter optimization and proxy re-encryption. A semi-supervised model of federated learning based on generative adversarial network is designed so that each participant can automatically adjust and optimize their own training parameters according to their own data distribution and environment in order to adapt to the data heterogeneity and dynamics of federated learning. Finally, experiments demonstrate that the proposed method can significantly improve the efficiency of data training and model performance compared with traditional federated learning methods across different datasets and scenarios. Experimental results show that the proposed strategy can achieve the same or even better accuracy than centralized machine learning while safeguarding data privacy and enabling ciphertext conversion. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Correction: Deep learning assists early-detection of hypertension-mediated heart change on ECG signals

引用

Hypertension research : official journal of the Japanese Society of Hypertension 2025年第2期48卷 874-875页

作者： Chengwei Liang Fan Yang Xiaobing Huang Lijuan Zhang Ying Wang Department of Automation Xiamen University Xiamen Fujian China. National Institute for Data Science in Health and Medicine Xiamen University Xiamen Fujian China. Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision Xiamen Fujian China. Fuzhou First General Hospital Fujian Medical University Fujian China. The First Affiliation Hospital of Xiamen University Xiamen University Xiamen Fujian China. zhangli.jennifer@***. Department of Automation Xiamen University Xiamen Fujian China. wangying@***. National Institute for Data Science in Health and Medicine Xiamen University Xiamen Fujian China. wangying@***. Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision Xiamen Fujian China. wangying@***.

来源：评论

学校读者我要写书评

暂无评论

NSPG-Miner: Mining Repetitive Negative Sequential Patterns

arXiv

引用

arXiv 2025年

作者： Li, Yan Wang, Zhulin Liu, Jing Guo, Lei Fournier-Viger, Philippe Wu, Youxi Wu, Xindong School of Economics and Management Hebei University of Technology Tianjin300401 China School of Artificial Intelligence Hebei University of Technology Tianjin300401 China State Key Laboratory of Reliability and Intelligence of Electrical Equipment Hebei University of Technology Tianjin300401 China College of Computer Science and Software Engineering Shenzhen University Shenzhen518061 China Hebei Key Laboratory of Big Data Computing 300401 China Key Laboratory of Knowledge Engineering with Big Data The Ministry of Education of China Hefei University of Technology Hefei230009 China

Sequential pattern mining (SPM) with gap constraints (or repetitive SPM or tandem repeat discovery in bioinformatics) can find frequent repetitive subsequences satisfying gap constraints, which are called positive sequential patterns with gap constraints (PSPGs). However, classical SPM with gap constraints cannot find the frequent missing items in the PSPGs. To tackle this issue, this paper explores negative sequential patterns with gap constraints (NSPGs). We propose an efficient NSPG-Miner algorithm that can mine both frequent PSPGs and NSPGs simultaneously. To effectively reduce candidate patterns, we propose a pattern join strategy with negative patterns which can generate both positive and negative candidate patterns at the same time. To calculate the support (frequency of occurrence) of a pattern in each sequence, we explore a NegPair algorithm that employs a key-value pair array structure to deal with the gap constraints and the negative items simultaneously and can avoid redundant rescanning of the original sequence, thus improving the efficiency of the algorithm. To report the performance of NSPG-Miner, 11 competitive algorithms and 11 datasets are employed. The experimental results not only validate the effectiveness of the strategies adopted by NSPG-Miner, but also verify that NSPG-Miner can discover more valuable information than the state-ofthe-art algorithms. Algorithms and datasets can be downloaded from https://***/wuc567/PatternMining/tree/master/NSPG-Miner. Copyright © 2025, The Authors. All rights reserved.

关键词： Miners

来源：评论

学校读者我要写书评

暂无评论

StyleGene: Crossover and Mutation of Region-level Facial Genes for Kinship Face Synthesis

StyleGene: Crossover and Mutation of Region-level Facial Gen...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Hao Li Xianxu Hou Zepeng Huang Linlin Shen Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen University School of AI and Advanced Computing Xi'an Jiaotong-Liverpool University Shenzhen Institute of Artificial Intelligence and Robotics for Society Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University

High-fidelity kinship face synthesis has many potential applications, such as kinship verification, missing child identification, and social media analysis. However, it is challenging to synthesize high-quality descendant faces with genetic relations due to the lack of large-scale, high-quality annotated kinship data. This paper proposes RFG (Region-level Facial Gene) extraction framework to address this issue. We propose to use IGE (Image-based Gene Encoder), LGE (Latent-based Gene Encoder) and Gene Decoder to learn the RFGs of a given face image, and the relationships between RFGs and the latent space of Style-GAN2. As cycle-like losses are designed to measure the $\mathcal{L}_data$ distances between the output of Gene Decoder and image encoder, and that between the output of LGE and IGE, only face images are required to train our framework, i.e. no paired kinship face data is required. Based upon the proposed RFGs, a crossover and mutation module is further designed to inherit the facial parts of parents. A Gene Pool has also been used to introduce the variations into the mutation of RFGs. The diversity of the faces of descendants can thus be significantly increased. Qualitative, quantitative, and subjective experiments on FIW, TSKinFace, and FF-databases clearly show that the quality and diversity of kinship faces generated by our approach are much better than the existing state-of-the-art methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FIRA: CAN WE ACHIEVE FULL-RANK TRAINING OF LLMS UNDER LOW-RANK CONSTRAINT?

arXiv

引用

arXiv 2024年

作者： Chen, Xi Feng, Kaituo Li, Changsheng Lai, Xunhao Yue, Xiangyu Yuan, Ye Wang, Guoren Beijing Institute of Technology China School of Intelligence Science and Technology Peking University China MMLab The Chinese University of Hong Kong Hong Kong Hebei Province Key Laboratory of Big Data Science and Intelligent Technology China

Low-rank training has emerged as a promising approach for reducing memory usage in training Large Language Models (LLMs). Previous methods either rely on decomposing weight matrices (e.g., LoRA), or seek to decompose gradient matrices (e.g., GaLore) to ensure reduced memory consumption. However, both of them constrain the training in a low-rank subspace, thus inevitably leading to sub-optimal performance. This raises a question: whether it is possible to consistently preserve the low-rank constraint for memory efficiency, while achieving full-rank training (i.e., training with full-rank gradients of full-rank weights) to avoid inferior outcomes? In this paper, we propose a new plug-and-play training framework for LLMs called Fira, as the first attempt to achieve this goal. First, we observe an interesting phenomenon during LLM training: the scaling impact of adaptive optimizers (e.g., Adam) on the gradient norm remains similar from low-rank to full-rank training. Based on this observation, we propose a norm-based scaling method, which utilizes the scaling impact of low-rank optimizers as substitutes for that of original full-rank optimizers to enable full-rank training. In this way, we can preserve the low-rank constraint in the optimizer while achieving full-rank training for better performance. Moreover, we find that there are sudden gradient rises during the optimization process, potentially causing loss spikes. To address this, we further put forward a norm-growth limiter to smooth the gradient via regulating the relative increase of gradient norms. Extensive experiments on the pre-training and fine-tuning of LLMs show that Fira outperforms both LoRA and GaLore, achieving performance that is comparable to or even better than full-rank training. For instance, our Fira can reduce the memory usage of optimizer states by 61.1%, while achieving improved performance for pre-training on the LLaMA 1B architecture. Notably, for pre-training on the LLaMA 7B architecture,

关键词： Memory architecture

来源：评论

学校读者我要写书评

暂无评论

Effective and Imperceptible Adversarial Textual Attack Via Multi-objectivization

引用

ACM Transactions on Evolutionary Learning and Optimization 2024年第3期4卷 1-23页

作者： Liu, Shengcai Lu, Ning Hong, Wenjing Qian, Chao Tang, Ke Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen College of Engineering South Tower 1088 Xueyuan Avenue Guangdong Shenzhen518055 China Department of Computer Science and Engineering Hong Kong University of Science and Technology Hong Kong999077 Hong Kong National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen 10F Zhizhen Building Canghai Campus 3688 Nanhai Avenue Guangdong Shenzhen China National Key Laboratory for Novel Software Technology School of Artificial Intelligence Nanjing University Nanjing International College Area Xianlin Campus 163 Xianlin Avenue Jiangsu Nanjing China

The field of adversarial textual attack has significantly grown over the past few years, where the commonly considered objective is to craft adversarial examples (AEs) that can successfully fool the target model. However, the imperceptibility of attacks, which is also essential for practical attackers, is often left out by previous studies. In consequence, the crafted AEs tend to have obvious structural and semantic differences from the original human-written text, making them easily perceptible. In this work, we advocate leveraging multi-objectivization to address such an issue. Specifically, we reformulate the problem of crafting AEs as a multi-objective optimization problem, where the attack imperceptibility is considered as an auxiliary objective. Then, we propose a simple yet effective evolutionary algorithm, dubbed HydraText, to solve this problem. HydraText can be effectively applied to both score-based and decision-based attack settings. Exhaustive experiments involving 44,237 instances demonstrate that HydraText consistently achieves competitive attack success rates and better attack imperceptibility than the recently proposed attack approaches. A human evaluation study also shows that the AEs crafted by HydraText are more indistinguishable from human-written text. Finally, these AEs exhibit good transferability and can bring notable robustness improvement to the target model by adversarial training. © 2024 Copyright held by the owner/author(s).

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

Exploring Patent Transformation Event: Forecasting Patent Transfer Time

Exploring Patent Transformation Event: Forecasting Patent Tr...

引用

International Conference on Computer Supported Cooperative Work in Design

作者： Jinchen Huo Weidong Liu Yang Li Yan Cao College of Computer Science Inner Mongolia University China Institute of Scientific and Technical Information of China China National & Local Joint Engineering Research Center of Intelligent Information Processing Technology for Mongolian China Inner Mongolia Key Laboratory of Social Computing and Data Processing China Inner Mongolia Engineering Laboratory for Big Data Analysis Technology China

As the largest source of technical information around the world, patents are regarded as an essential crystallization and carrier of knowledge and technological innovation. Patent transformation is conducive not only to enhancing economic efficiency, but also to improving productivity and the rational utilization of resources. There is an imbalance between high patent ownership and low transformation rates. We try to predict the occurrence of transformation events from the patent assignment. However, there are some challenges in predicting patent transformation: (1) how to capture transformation features of patents, especially combined with the transfer time factor. (2) how to predict patent transfer time effectively. To address these challenges, a Patent Transfer Time Forecasting Model (PTTFM) is proposed. The model includes: (1) extraction of time-varying features of patents. (2) the patent transfer time is forecast using a Neural Temporal Point Process. By testing the model on patents under different classifications, the experimental results are obtained to show that the proposed model is applicable to predict the timing of patent assignment within a certain time frame, especially one month. Our work may facilitate patent transformation while interpretability is ensured for transformation events.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lgccn: A Label-Guided Contrastive Capsule Network for Few-Shot Text Classification

SSRN

引用

SSRN 2023年

作者： Xu, Yang Xu, Xiaolong Xiao, Fu Jiangsu Key Laboratory of Big Data Security & Intelligent Processing Nanjing University of Posts and Telecommunications Nanjing China School of Computer Science Nanjing University of Posts and Telecommunications Nanjing China

Few-shot text classification aims to learn transferable knowledge from a limited dataset to perform classification tasks in unseen domains. Recently, metric-based meta-learning methods have demonstrated their advantages in few-shot text classification. However, the influence of randomly sampling support set on prototype quality has often been overlooked by these methods. In this paper, we propose a Label-Guided Contrastive Capsule Network (LGCCN) to address this issue. LGCCN leverages label information to guide supervised contrastive learning, using labels as anchors to enhance the clustering process, resulting in higher-quality sentence vectors. Furthermore, LGCCN employs label-guided dynamic routing process to iteratively generate a prototype, significantly mitigating the impact of random support set sampling. We conducted extensive experiments on four datasets, comparing LGCCN with recent methods. The results reveal that LGCCN, compared to the recent best work, has achieved an improvement of 10.1% and 2.1% in classification accuracy for 1-shot and 5-shot tasks, respectively. Additionally, ablation studies validate the effectiveness of each module in our approach. © 2023, The Authors. All rights reserved.

关键词： Iterative methods

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：