检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Wang, Hongjun Chen, Jiyuan Pan, Tong Dong, Zheng Zhang, Lingyu Jiang, Renhe Song, Xuan SUSTech-UTokyo Joint Research Center on Super Smart City Department of Computer Science and Engineering Shenzhen China Shenzhen China School of Artificial Intelligence Jilin Universit Shenzhen China Center for Spatial Information Science University of Tokyo Tokyo Japan

Traffic forecasting is a cornerstone of smart city management, enabling efficient resource allocation and transportation planning. Deep learning, with its ability to capture complex nonlinear patterns in spatiotemporal (ST) data, has emerged as a powerful tool for traffic forecasting. While graph neural networks (GCNs) and transformer-based models have shown promise, their computational demands often hinder their application to real-world road networks, particularly those with large-scale spatiotemporal interactions. To address these challenges, we propose a novel spatiotemporal graph transformer (STGformer) architecture. STGformer effectively balances the strengths of GCNs and Transformers, enabling efficient modeling of both global and local traffic patterns while maintaining a manageable computational footprint. Unlike traditional approaches that require multiple attention layers, STG attention block captures high-order spatiotemporal interactions in a single layer, significantly reducing computational cost. In particular, STGformer achieves a 100x speedup and a 99.8% reduction in GPU memory usage compared to STAEformer during batch inference on a California road graph with 8,600 sensors. We evaluate STGformer on the LargeST benchmark and demonstrate its superiority over state-of-the-art Transformer-based methods such as PDFormer and STAEformer, which underline STGformer’s potential to revolutionize traffic forecasting by overcoming the computational and memory limitations of existing approaches, making it a promising foundation for future spatiotemporal modeling tasks. Codes are available at GitHub © 2024, CC BY.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Joint Sensing, Communication and Computation for Status Update in Mobile Edge Computing with Non-Orthogonal Multiple Access

引用

IEEE Internet of Things Journal 2025年

作者： Zhang, Jianfei Wang, Yufan Hu, Yun Chang, Zheng Qiqihar University College of Computer and Control Engineering Qiqihar161003 China Xidian University National Demonstration Center for Experimental Electronic Information and Communications Engineering Education Xi’an710071 China University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu611731 China University of Jyväskylä Faculty of Information Technology Jyväskylä40014 Finland

Mobile Edge Computing (MEC) is considered as a promising solution for augmenting the computational capabilities of Internet of Things (IoT) devices by offloading tasks to nearby edge servers (ESs). However, finite communication and computational resources at both IoT devices and ESs, coupled with escalating congestion as more devices connect, present critical challenges in maintaining low latency and high efficiency. In this work, we jointly design the task sensing, communication, and resource allocation for MEC with Non-Orthogonal Multiple Access (NOMA). To address the urgent need for timely data processing in IoT applications, we utilize the Age of Information (AoI) metric as a measure of data freshness. With the objective to minimize the system cost, we propose to jointly optimize sensing sampling intervals, sensing frequencies, offloading decision, and power allocation. Recognizing that this problem is NP-hard, we decompose it incrementally and propose a High-Dimensional Progressive Cost Optimization (HDPCO) algorithm to reduce overall system cost. Simulation results confirm the effectiveness of HDPCO, showing significant improvements in minimizing overall system cost compared to other proposed schemes. © 2014 IEEE.

关键词： Data reduction

来源：评论

学校读者我要写书评

暂无评论

SigStyle: Signature Style Transfer via Personalized Text-to-Image Models 39

SigStyle: Signature Style Transfer via Personalized Text-to-...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Wang, Ye Bai, Tongyuan Xie, Xuping Yi, Zili Wang, Yilin Ma, Rui School of Artificial Intelligence Jilin University China College of Computer Science and Technology Jilin University China School of Intelligence Science and Technology Nanjing University China Adobe United States Engineering Research Center of Knowledge-Driven Human-Machine Intelligence MOE China

ISBN: (纸本)157735897X

Style transfer enables the seamless integration of artistic styles from a style image into a content image, resulting in visually striking and aesthetically enriched outputs. Despite numerous advances in this field, existing methods did not explicitly focus on the signature style, which represents the distinct and recognizable visual traits of the image such as geometric and structural patterns, color palettes and brush strokes etc. In this paper, we introduce SigStyle, a framework that leverages the semantic priors that embedded in a personalized text-to-image diffusion model to capture the signature style representation. This style capture process is powered by a hypernetwork that efficiently fine-tunes the diffusion model for any given single style image. Style transfer then is conceptualized as the reconstruction process of content image through learned style tokens from the personalized diffusion model. Additionally, to ensure the content consistency throughout the style transfer process, we introduce a time-aware attention swapping technique that incorporates content information from the original image into the early denoising steps of target image generation. Beyond enabling high-quality signature style transfer across a wide range of styles, SigStyle supports multiple interesting applications, such as local style transfer, texture transfer, style fusion and style-guided text-to-image generation. Quantitative and qualitative evaluations demonstrate our approach outperforms existing style transfer methods for recognizing and transferring the signature styles. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Hypertext systems

来源：评论

学校读者我要写书评

暂无评论

LE-CAM++: A Lighter and More Efficient CAM++ for Speaker Verification

LE-CAM++: A Lighter and More Efficient CAM++ for Speaker Ver...

引用

International Symposium on Chinese Spoken Language Processing

作者： Shuanghong Liu Zhida Song Zhihua Fang Liang He School of Computer Science and Technology Xinjiang University Urumqi Xinjiang Key Laboratory of Signal Detection and Processing Urumqi School of Intelligence Science and Technology Xinjiang University Urumqi Department of Electronic Engineering and Beijing National Research Center for Information Science and Technology Tsinghua University Beijing

ISBN: (数字)9798331516826

ISBN: (纸本)9798331516833

Due to its superior performance and fewer parameters, CAM++ has become the state-of-the-art model for speaker verification tasks. This model uses 2D convolutional blocks to extract front-end features, which are then fed into a densely connected time-delay neural network backbone to extract deep features. However, the simple stacking of 2D convolutions may lead to the generation of a significant amount of redundant features, which is detrimental to efficient feature extraction. Furthermore, although CAM++ already has a relatively small number of parameters, there is still room for further optimization. To address these issues, this paper first employs depthwise separable convolutions to replace the dilated convolutions in the back-end network of CAM++, making the model more lightweight. Next, we introduce spatial and channel reconstruction convolution (SCConv) in the ResBlock module of CAM++ to reduce redundant features and optimize the feature extraction process. Finally, after SCConv, we apply squeeze and excitation attention mechanism to model the interdependencies between channels and recalibrate each channel, further enhancing the model's representational capacity. We name the resulting model LE-CAM++. Our proposed model achieves an EER of 0.686 and a minDCF of 0.084 on the VoxCeleb1–O dataset. Compared to the baseline model CAM ++, the EER is reduced by 11%, and the minDCF is reduced by 28%. Additionally, the model parameters are reduced by 8%.

关键词： Attention mechanisms Stacking Neural networks Feature extraction Optimization

来源：评论

学校读者我要写书评

暂无评论

EventAid: Benchmarking Event-aided Image/Video Enhancement Algorithms with Real-captured Hybrid Dataset

arXiv

引用

arXiv 2023年

作者： Duan, Peiqi Li, Boyu Yang, Yixin Lou, Hanyue Teng, Minggui Ma, Yi Shi, Boxin National Key Laboratory for Multimedia Information Processing National Engineering Research Center of Visual Technology School of Computer Science Peking University China

—Event cameras are emerging imaging technology that offers advantages over conventional frame-based imaging sensors in dynamic range and sensing speed. Complementing the rich texture and color perception of traditional image frames, the hybrid camera system of event and frame-based cameras enables high-performance imaging. With the assistance of event cameras, high-quality image/video enhancement methods make it possible to break the limits of traditional frame-based cameras, especially exposure time, resolution, dynamic range, and frame rate limits. This paper focuses on five event-aided image and video enhancement tasks (i.e., event-based video reconstruction, event-aided high frame rate video reconstruction, image deblurring, image super-resolution, and high dynamic range image reconstruction), provides an analysis of the effects of different event properties, a real-captured and ground truth labeled benchmark dataset, a unified benchmarking of state-of-the-art methods, and an evaluation for two mainstream event simulators. In detail, this paper collects a real-captured evaluation dataset EVENTAID for five event-aided image/video enhancement tasks, by using "Event-RGB" multi-camera hybrid system, taking into account scene diversity and spatiotemporal synchronization. We further perform quantitative and visual comparisons for state-of-the-art algorithms, provide a controlled experiment to analyze the performance limit of event-aided image deblurring methods, and discuss open problems to inspire future research. Copyright © 2023, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

High-Performance Asynchronous CNN Accelerator with Early Termination 15

High-Performance Asynchronous CNN Accelerator with Early Ter...

引用

15th IEEE International Symposium on Embedded Multicore/Many-Core Systems-on-Chip, MCSoC 2022

作者： Loo, Tan Rong Teo, T. Hui Tiruye, Mulat Ayinet Wey, I-Chyn Singapore University of Technology and Design Engineering Product Development Singapore Singapore University of Technology and Design Engineering Product Development Science Mathematics and Technology Singapore Artificial Intelligence Research Center School of Electrical and Computer Engineering College of Engineering Graduate Institute of Electrical Engineering Chang Gung University Chang Gung University Taiwan

ISBN: (纸本)9781665464994

Convolutional Neural Network (CNN), especially very deep networks, are highly computation intensive, resulting in long delays and high-power consumption. The dynamically varying environmental conditions in real world inference can result in highly complex problems, and hence the need for these inefficient deep networks to guarantee satisfactory accuracy all the time. Several studies employ approximation techniques to execute partial computation of the network, in attempt to reduce the amount of computation were unnecessary. However, such approaches are still highly sequential in nature, since they still need to run the whole network. This paper proposes an early termination architecture on an already-trained CNN to allow for testing the partial results midway through the network, reducing computations by terminating the main network when it is sufficient as the inference results. The first proposal is implemented in synchronous circuit, however, due to its nature all memory elements are required to capture even when no new data is generated. The second proposal employs the use of asynchronous circuit to significantly reducing power consumption and further sped up the architecture since an operation need not wait the slowest critical path in the circuit. The proposed circuits were designed on FPGA platform. The results of the asynchronous circuit show a nearly 20% increment in speed with about 12% reduction in power consumption in comparison with a synchronous circuit. © 2022 IEEE.

关键词： Electric power utilization

来源：评论

学校读者我要写书评

暂无评论

Doping Shortens the Metal/Metal Distance and Promotes OH Coverage in Non-Noble Acidic Oxygen Evolution Reaction Catalysts

引用

Journal of the American Chemical Society 2023年第14期145卷 7829-7836页

作者： Wang, Ning Ou, Pengfei Miao, Rui Kai Chang, Yuxin Wang, Ziyun Hung, Sung-Fu Abed, Jehad Ozden, Adnan Chen, Hsuan-Yu Wu, Heng-Liang Huang, Jianan Erick Zhou, Daojin Ni, Weiyan Fan, Lizhou Yan, Yu Peng, Tao Sinton, David Liu, Yongchang Liang, Hongyan Sargent, Edward H. Department of Electrical and Computer Engineering University of Toronto 35 St George Street TorontoONM5S 1A4 Canada School of Materials Science and Engineering Key Laboratory of Efficient Utilization of Low and Medium Grade Energy Ministry of Education Tianjin University Tianjin300350 China Department of Mechanical and Industrial Engineering University of Toronto 5 King’s College Road TorontoONM5S 3G8 Canada School of Chemical Sciences The University of Auckl Auckland1010 New Zealand Department of Applied Chemistry National Yang Ming Chiao Tung University Hsinchu300 Taiwan Center for Condensed Matter Sciences National Taiwan University Taipei10617 Taiwan Center of Atomic Initiative for New Materials National Taiwan University Taipei10617 Taiwan State Key Lab of Hydraulic Engineering Simulation and Safety School of Materials Science & Engineering Tianjin University Tianjin300354 China

Acidic water electrolysis enables the production of hydrogen for use as a chemical and as a fuel. The acidic environment hinders water electrolysis on non-noble catalysts, a result of the sluggish kinetics associated with the adsorbate evolution mechanism, reliant as it is on four concerted proton-electron transfer steps. Enabling a faster mechanism with non-noble catalysts will help to further advance acidic water electrolysis. Here, we report evidence that doping Ba cations into a Co3O4 framework to form Co3-xBaxO4 promotes the oxide path mechanism and simultaneously improves activity in acidic electrolytes. Co3-xBaxO4 catalysts reported herein exhibit an overpotential of 278 mV at 10 mA/cm2 in 0.5 M H2SO4 electrolyte and are stable over 110 h of continuous water oxidation operation. We find that the incorporation of Ba cations shortens the Co-Co distance and promotes OH adsorption, findings we link to improved water oxidation in acidic electrolyte. © 2023 American Chemical Society.

关键词： Catalysts

来源：评论

学校读者我要写书评

暂无评论

RASLS: Reinforcement Learning Active SLAM Approach with Layout Semantic

RASLS: Reinforcement Learning Active SLAM Approach with Layo...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Changhang Tian Shasha Tian Yilin Kang Hong Wang Jun Tie Shenhao Xu School of Computer Science South-Central Minzu University Wuhan China Hubei Provincial Engineering Research Center for Intelligent Management of Manufacturing Enterprise Wuhan China Hubei Provincial Engineering Research Center of Agricultural Blockchain and Intelligent Management Wuhan China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Active SLAM plays a crucial role in applications of embodied intelligence. Previous learning-based methods struggle to sufficiently leverage semantic information in the environment, while frontier-based algorithms face challenges in mitigating myopic decision-making issues. Considering that humans experientially use observed information while exploring environments, we propose a deep reinforcement learning approach incorporating object semantic information and design a reward-matching mechanism based on the prior object layout. To tackle the instability in exploration gains caused by error optimization in the mapping part of the SLAM system, we introduce a method for differential map uncertainty confidence filtering. We conduct reinforcement learning training using Gazebo in office scenarios based on 3DGEMS and perform comparisons in a new scenario. Through ablation analysis, we demonstrate the effectiveness of layout semantic information. Compared to the latest reinforcement learning baseline, experimental results indicate that our method achieves a higher success rate with a shorter average execution time and path length. Our code and training world is available at https://***/Moresweet/DRL_ARE_Graph_Gazebo.

关键词： Training Location awareness Simultaneous localization and mapping Uncertainty Semantics Layout Trajectory

来源：评论

学校读者我要写书评

暂无评论

MedSeq: Semantic Segmentation for Medical Image Sequences

MedSeq: Semantic Segmentation for Medical Image Sequences

引用

2022 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022

作者： Yuan, Runtian Xu, Jilan Li, Xiaotong Zhang, Yuejie Feng, Rui Zhang, Xiaobo Zhang, Tao Gao, Shang Fudan University School of Computer Science Shanghai China Children's Hospital of Fudan University National Children's Medical Center China Shanghai University of Finance and Economics School of Information Management and Engineering Shanghai China Deakin University School of Information Technology Victoria Australia

ISBN: (纸本)9781665468190

Medical image segmentation plays a critical role in computer-aided diagnosis, while the diversity and complexity of medical images make it difficult to segment precisely. In practice, medical images of specific modalities (e.g. Magnetic Resonance Imaging, Colonoscopy and Ultrasonography) are collected as sequences independently for every patient. However, 1) there exists few works exploiting sequence information among successive frames, neglecting inter-frame relationships that are useful to locate target objects;2) the performance of medical image segmentation is limited to the low contrast or blurry boundary of medical images, and intra-frame dependencies are not fully explored. Thus in this paper, we propose MedSeq for segmenting objects of interest in medical image sequences. Following the 'locate-then-refine' paradigm, we locate target regions by modeling cross-frame relationships and then perform refinement on coarse masks. More specifically, we design a Cross-frame Attention module to learn correlations among frames, taking advantages of their similar appearances. For refinement, we propose a novel Boundary-aware Transformer to improve the segmentation of boundary patches. Extensive experiments are conducted on benchmark datasets of Cardiac Segmentation and Video Polyp Segmentation. Our method achieves superior performance over the state-of-the-art methods. © 2022 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Imperceptible adversarial attack with entropy feature and segmentation-based constraint 2021

Imperceptible adversarial attack with entropy feature and se...

引用

2021 10th International Conference on Computing and Pattern Recognition

作者： Rongdong Li Qinliang Lin Yinglong Fu Weicheng Xie Linlin Shen School of Computer Science&Software Engineering Shenzhen University China and Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen University China

ISBN: (纸本)9781450390439

Methods of adversarial attack and defense have attracting increasing attention in the fields of security and protection related applications. However, current algorithms carry out perturbations on entire images and mostly consider their imperceptibility to machines, while does not take their human imperceptibility into account. In this work, we propose a constrained adversarial attack algorithm with both machine and human imperceptibility based on image entropy feature and accurate segmentation. The proposed algorithm has three merits. First, image entropy-based feature for quantifying the imperceptibility of a semantic region is introduced, which is simple yet efficient to implement. Second, in terms of the imperceptibility metric, accurate target regions for adversarial perturbation are obtained based on scene-aware segmentation and merging. Third, a general adversarial attack based on segmentation region constraint is proposed to induce both machine and visual imperceptibility. Experimental results in terms of qualitative and quantitative analysis reflect the effectiveness of the proposed algorithm compared with the state of the arts.

关键词： Adversarial attack semantic segmentation robust object classification imperceptibility metric constrained attack algorithm

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：