检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Zheng Ti Xuze Hao Renhai Chen College of Intelligence and Computing Tianjin University Tianjin China School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Fudan University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Continual Semantic Segmentation (CSS) aims to continuously learn new classes while mitigating catastrophic forgetting. Existing CSS methods primarily address this challenge through knowledge distillation. While they focus on maintaining stability for old classes, this emphasis often restricts plasticity in learning new ones. In this work, we propose two plug-and-play methods to better balance stability and plasticity. First, for feature map distillation, we replace pixel-level constraints with statistical information extraction within a window, improving the learning of new classes while preserving structural integrity to overcome forgetting. Second, we use relative ranks of predictions to replace exact probabilistic values to loosen the constraints in the logit distribution of old classes. Extensive experiments demonstrate that our methods outperform state-of-the-art methods when integrating into several baselines across various CSS scenarios.

关键词： Semantic segmentation Semantics Signal processing Feature extraction Probabilistic logic information retrieval Acoustics Data mining Speech processing

来源：评论

学校读者我要写书评

暂无评论

Dual-layer Broadband High-gain Circularly Polarized Patch Antenna with L-shaped Stub

Dual-layer Broadband High-gain Circularly Polarized Patch An...

引用

2023 IEEE International Workshop on Electromagnetics: Applications and Student Innovation Competition, iWEM 2023

作者： Li, Wen Xue, Wei Li, Yingsong Loh, Tian Hong Chung, Kwok L. Harbin Engineering University College of Information and Communication Engineering Harbin150001 China Anhui University The Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Hefei China National Physical Laboratory Electromagnetic and Eletrochemical Technologies Department Middlesex United Kingdom Huizhou University School of Computing Science and Engineering Huizhou516007 China

ISBN: (纸本)9798350336740

This paper introduced a dual-layer broadband high-gain circularly polarized (CP) antenna. The proposed broadband CP antenna consists of the radiating patch and couping structure, which is made up of opening slot patch and L-shaped stub. An axial-ratio bandwidth of 26 % (4.36-5.66 GHz) is shown by this proposed dual-layer CP patch antenna with a simulated -10-dB reflection coefficient bandwidth of 37 % (4.23-6.15 GHz). The broadband characteristics offered by this CP antenna are promising for a wide range of future applications. © 2023 IEEE.

关键词： Microstrip antennas

来源：评论

学校读者我要写书评

暂无评论

An Inversion-free Fuzzy Zeroing Neural Network Under Adaptive Input Range Fuzzy Scheme: Design, Analysis, and Application

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年

作者： Xiao, Lin Wang, Dan Zuo, Qiuyue Fan, Haibing Cai, Hang Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Changsha410081 China Hunan Normal University MOE-LCSM Changsha410081 China

This article proposes a predefined-time robust inversion-free fuzzy zeroing neural network (PRIFZNN) model, which comprehensively considers computational complexity, convergence, and robustness. The PRIFZNN model achieves the lowest computational complexity among zeroing neural network (ZNN) models for solving similar problems. By avoiding the matrix inversion operation, it reduces the computational complexity to O(mn), whereas other ZNN models typically exhibit a complexity level of O(min(mn2,m2n)). To enhance the convergence and robustness of the PRIFZNN model, a piecewise predefined-time robust activation function and a novel adaptive input range fuzzy control scheme are proposed. The former provides predefined-time convergence and the ability to resist bounded noise, while the latter is employed to further improve the convergence rate. Moreover, the theoretical analysis section offers a detailed discussion of the convergence properties of the PRIFZNN model. It also derives the upper bound of the convergence time under bounded noise interference, thereby theoretically ensuring the robustness of the PRIFZNN model. Finally, numerical simulation studies on solving the time-varying Sylvester matrix equation show that the PRIFZNN model surpasses the comparative ZNN models in both convergence and robustness, while also validating the correctness of the theoretical analysis. Meanwhile, the PRIFZNN model is successfully applied to the path-tracking tasks of robotic manipulators, demonstrating its excellent practicality. © 2020 IEEE.

关键词： NP-hard

来源：评论

学校读者我要写书评

暂无评论

Diverse and High-Quality Food Image Generation from Only Food Names

引用

ACM Transactions on Multimedia computing, Communications and Applications 2025年第5期21卷

作者： Yu, Dongjian Min, Weiqing Jin, Xin Jiang, Qian Jin, Ying Jiang, Shuqiang School of Software Yunnan University Kunming China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences University of Chinese Academy of Sciences Beijing China

Food image generation holds promising application prospects in food design, advertising, and food education. However, the existing methods rely on information such as recipes, ingredients, or food names, which leads to generated food images with less intra-class diversity. When recipes, ingredients, and food names are identical for the same food, the real-world images may vary significantly in appearance. The question of how to simultaneously ensure the quality and diversity of the generated images is a key issue. To this end, we employ pre-trained diffusion model and Transformer to propose a method for generating diverse and high-quality images of both Chinese and Western food, named CW-Food. Different from previous works that utilize an overall food feature to generate new images, CW-Food first decouples the food images to obtain common intra-class features and private instance features. Additionally, we design a Transformer-based feature fusion module to integrate the common and private features, in order to avoid the shortcomings of conventional methods. Moreover, we also utilize a pre-trained diffusion model as our backbone, which is fine-tuned using LoRA with the fused multi-variate features. Extensive experiments on four datasets demonstrate the advantages of our proposed method, producing diverse and high-quality food images encompassing both Chinese and Western cuisines. To the best of our knowledge, our work is the first attempt to generate Chinese food images using only food names. © 2025 Copyright held by the owner/author(s).

关键词： Food ingredients

来源：评论

学校读者我要写书评

暂无评论

A preconditioned iteration method for solving saddle point problems

arXiv

引用

arXiv 2024年

作者： Zhang, Juan Luo, Yiyi Hunan Key Laboratory for Computation and Simulation in Science and Engineering Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education School of Mathematics and Computational Science Xiangtan University China

This paper introduces a preconditioned method designed to comprehensively address the saddle point system with the aim of improving convergence efficiency. In the preprocessor construction phase, a technical approach for solving the approximate inverse matrix of sparse matrices is presented. The effectiveness of the proposed method is demonstrated through numerical examples, emphasizing its efficacy in approximating the inverse matrix. Furthermore, the preprocessing technology includes a low-rank processing step, effectively reducing algorithmic complexity. Numerical experiments validate the effectiveness and feasibility of PSLR-GMRES in solving the saddle point system. © 2024, CC BY.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Blockchain-based Transaction Scheme for Massive MIMO Channel Estimation 11

Blockchain-based Transaction Scheme for Massive MIMO Channel...

引用

11th International Conference on Networks, Communication and computing, ICNCC 2022

作者： Huang, Yu Zhi, Hui Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education Anhui University Hefei230601 China School of Electronic and Information Engineering Anhui University Hefei230601 China

ISBN: (纸本)9781450398039

In wireless communication systems, accurate channel estimation is essential to ensure the performance of wireless communication systems. Massive Multiple Input Multiple Output (M-MIMO) systems have a dramatic increase in the number of channels to be estimated due to the large number of antennas, which poses a great challenge to the computing resources of the base station, and one solution is offloading these computing tasks of channel estimation to other computing nodes with more computing resources (such as servers). However, in real life, computing nodes are not willing to participate in channel estimation because of their equipment and operation cost. In order to solve this problem, this paper introduces blockchain technology and designs a blockchain-based transaction scheme for channel estimation (BTS-CE) to solve the problem of computing resources transaction for massive MIMO channel estimation, and designs the specific process of the scheme. In order to solve the problem that computing nodes cannot complete the computation task due to failure or malicious disconnection, a task allocation model is designed to assign a computation task to multiple computation nodes. The security of the transactions is guaranteed through the blockchain. Practical Byzantine Fault Tolerance (PBFT) consensus mechanism cannot be applied to the public chain due to its low performance caused by a large number of nodes participating in consensus, this problem is solved by selecting some active nodes as consensus nodes. Simulation results show that improved PBFT(IPBFT) is lower in transaction delay and higher in throughput than PoW consensus mechanism. © 2022 ACM.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-based AP Selection Method in Software-defined Networks 5

Deep Learning-based AP Selection Method in Software-defined ...

引用

5th IEEE International Conference on Electronic information and Communication Technology, ICEICT 2022

作者： Duan, Miaomiao Zhi, Hui Yang, Lixia Anhui University Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Hefei230601 China School of Electronic and Information Engineering Anhui University Hefei230601 China

ISBN: (数字)9781665472111

ISBN: (纸本)9781665472111

In the Software Defined Network (SDN), the terminal has been connected to the network in the process when an access point failure or damage, if the access point is not properly selected, it will lead to a decline in the quality of service applications. To address this problem, most traditional AP selection methods use access point selection based on received signal strength without take into account the channel and channel capacity of the access point access, while this paper uses Deep Neural Networks (DNN)-based access point (AP) selection method and considers the channel and channel capacity as part of the parameters. In the SDN controller, the received signal strength, throughput, number of connected devices, channel, and the channel capacity and access point performance quality are used as the input and label of the DNN, respectively. The analysis and simulation results show that the algorithm can select AP with better performance quality compared with the traditional AP selection method based on received signal strength, and the correct rate is significantly improved by 5% when compared with the AP selection method based on feedforward neural network. Meanwhile, the DNN-based access point selection method in SDN achieves flexible control of network traffic and balances the load of access points in the network. © 2022 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Comments and extensions on "State-equivalent form and minimum-order compensator design for rectangular descriptor systems"

arXiv

引用

arXiv 2025年

作者： Shi, Shuo Zhang, Juan Hunan Key Laboratory for Computation and Simulation in Science and Engineering Key Laboratory of Intelligent Computing and Information Processing Ministry of Education School of Mathematics and Computational Science Xiangtan University Xiangtan411105 China

This technical note presents a counterexample showing that the equivalence conditions proposed by Geng et al. (IEEE Trans. Automat. Control, 2024), which use a minimum-order compensator (MOC) to achieve desired designs, including generalized regularity or both generalized regularity and free of impulse, are sufficient but not necessary. Furthermore, revised equivalence conditions are introduced, along with an equivalence condition ensuring the closed-loop system remains generalized regular, impulse-free, and stable using MOC. Additionally, it is shown that output feedback can replace the MOC, achieving the same design without increasing dimensionality. These findings are validated through a circuit *** Codes 93B05, 93B52, 15A03 Copyright © 2025, The Authors. All rights reserved.

关键词： Feedback

来源：评论

学校读者我要写书评

暂无评论

Y-Mol: A Multiscale Biomedical Knowledge-Guided Large Language Model for Drug Development

arXiv

引用

arXiv 2024年

作者： Ma, Tengfei Lin, Xuan Li, Tianle Li, Chaoyi Chen, Long Zhou, Peng Cai, Xibao Yang, Xinyu Zeng, Daojian Cao, Dongsheng Zeng, Xiangxiang College of Computer Science and Electronic Engineering Hunan University China Key Laboratory of Intelligent Computing and Information Processing Ministry of Education Xiangtan University China Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Hunan Normal University China Xiangya School of Pharmaceutical Sciences Central South University China

Large Language Models (LLMs) have recently demonstrated remarkable performance in general tasks across various fields. However, their effectiveness within specific domains such as drug development remains challenges. To solve these challenges, we introduce Y-Mol, forming a well-established LLM paradigm for the flow of drug development. Y-Mol is a multiscale biomedical knowledge-guided LLM designed to accomplish tasks across lead compound discovery, pre-clinic, and clinic prediction. By integrating millions of multiscale biomedical knowledge and using LLaMA2 as the base LLM, Y-Mol augments the reasoning capability in the biomedical domain by learning from a corpus of publications, knowledge graphs, and expert-designed synthetic data. The capability is further enriched with three types of drug-oriented instructions: description-based prompts from processed publications, semantic-based prompts for extracting associations from knowledge graphs, and template-based prompts for understanding expert knowledge from biomedical tools. Besides, Y-Mol offers a set of LLM paradigms that can autonomously execute the downstream tasks across the entire process of drug development, including virtual screening, drug design, pharmacological properties prediction, and drug-related interaction prediction. Our extensive evaluations of various biomedical sources demonstrate that Y-Mol significantly outperforms general-purpose LLMs in discovering lead compounds, predicting molecular properties, and identifying drug interaction events. The source code is available at https://***/r/Y-Mol. © 2024, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Open-vocabulary 3D Semantic Understanding via Affinity Neural Radiance Fields

Open-vocabulary 3D Semantic Understanding via Affinity Neura...

引用

Next Generation Data-driven Networks (NGDN), International Conference on

作者： Yujie Fan Huan Luo Computer and Data Science Fuzhou University Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University Fuzhou China

ISBN: (数字)9798350388374

ISBN: (纸本)9798350388381

Three-dimensional semantic understanding using only several multi-view images can largely reduce the communication burden on the network. In addition, while point clouds are extensively studied for 3D scene understanding, utilization of multi-view image data offers rich visual details and texture information. However, challenges persist in lifting 2D semantic features to 3D space and leveraging language for segmentation. Inspired by recent advancements, this paper proposes a method that combines CLIP features and SAM masks to create a feature field capable of segmenting objects via natural language text across 2D multi-view and 3D Gaussian splatting. It offers a promising function for extracting 3D assets for game engines and the metaverse. Our method involves mask generation from video frames, extracting physical scales via RGB Nerf with masks, and organizing hierarchical information for semantic comprehension. In the training process, affinity features maintain scale properties and guide CLIP feature generation with auto weights blending for semantic robustness. A straightforward 3D splatting CLIP feature approach and canonical text methodology enhance query robustness across 2D multi-view and 3D splatting through relevance score calculation based on text CLIP features for inference. Experimental results demonstrate promising improvements in semantic understanding of 3D scenes.

关键词： Training Image segmentation Vocabulary Visualization Three-dimensional displays Semantics Natural languages Neural radiance field Feature extraction Robustness

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：