检索结果-内蒙古大学图书馆

10th International Conference on Learning Representations, ICLR 2022

作者： Chen, Yixuan Shi, Yubin Li, Dongsheng Wang, Yujiang Dong, Mingzhi Zhao, Yingying Dick, Robert Lv, Qin Yang, Fan Shang, Li China and Shanghai Key Laboratory of Data Science School of Computer Science Fudan University Shanghai China Microsoft Research Asia Shanghai China Department of Computing Imperial College London London United Kingdom Department of Electrical Engineering and Computer Science University of Michigan Michigan United States Department of Computer Science University of Colorado Boulder Boulder United States School of Microelectronics Fudan University Shanghai China

Disentangled feature representation is essential for data-efficient learning. The feature space of deep models is inherently compositional. Existing β-VAE-based methods, which only apply disentanglement regularization to the resulting embedding space of deep models, cannot effectively regularize such compositional feature space, resulting in unsatisfactory disentangled results. In this paper, we formulate the compositional disentanglement learning problem from an information-theoretic perspective and propose a recursive disentanglement network (RecurD) that propagates regulatory inductive bias recursively across the compositional feature space during disentangled representation learning. Experimental studies demonstrate that RecurD outperforms β-VAE and several of its state-of-the-art variants on disentangled representation learning and enables more data-efficient downstream machine learning tasks. © 2022 ICLR 2022 - 10th International Conference on Learning Representationss. All rights reserved.

关键词： Information theory

来源：评论

学校读者我要写书评

暂无评论

DAMPER: A Dual-Stage Medical Report Generation Framework with Coarse-Grained MeSH Alignment and Fine-Grained Hypergraph Matching

arXiv

引用

arXiv 2024年

作者： Huang, Xiaofei Chen, Wenting Liu, Jie Lu, Qisheng Luo, Xiaoling Shen, Linlin Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen China Department of Electrical Engineering City University of Hong Kong Kowloon Hong Kong National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China

Medical report generation is crucial for clinical diagnosis and patient management, summarizing diagnoses and recommendations based on medical imaging. However, existing work often overlook the clinical pipeline involved in report writing, where physicians typically conduct an initial quick review followed by a detailed examination. Moreover, current alignment methods may lead to misaligned relationships. To address these issues, we propose DAMPER, a dual-stage framework for medical report generation that mimics the clinical pipeline of report writing in two stages. In the first stage, a MeSH-Guided Coarse-Grained Alignment (MCG) stage that aligns chest X-ray (CXR) image features with medical subject headings (MeSH) features to generate a rough keyphrase representation of the overall impression. In the second stage, a Hypergraph-Enhanced Fine-Grained Alignment (HFG) stage that constructs hypergraphs for image patches and report annotations, modeling high-order relationships within each modality and performing hypergraph matching to capture semantic correlations between image regions and textual phrases. Finally,the coarse-grained visual features, generated MeSH representations, and visual hypergraph features are fed into a report decoder to produce the final medical report. Extensive experiments on public datasets demonstrate the effectiveness of DAMPER in generating comprehensive and accurate medical reports, outperforming state-of-the-art methods across various evaluation metrics. Copyright © 2024, The Authors. All rights reserved.

关键词： Coarse-grained modeling

来源：评论

学校读者我要写书评

暂无评论

DiffSenseNet: Integrating Hierarchical Features and Angular Diffusion for Remote Sensing Object Detection

DiffSenseNet: Integrating Hierarchical Features and Angular ...

引用

International Symposium on Parallel and Distributed Processing with Applications, ISPA

作者： Songwen Pei Hongli Ma Xinyun Qiu Libo Huang Key Laboratory of Fire Detection and Early Warning Technology of Shanghai Emergency Management Bureau School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai China School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai China National University of Defense Technology Changsha China

ISBN: (数字)9798331509712

ISBN: (纸本)9798331509729

Object detection in remote sensing images is a significant and challenging task. The detection performance is difficult to further improve due to the wide variation in object scale and unpredictable orientations. However, traditional methods typically rely on Region Proposal Networks (RPN) with fixed anchor boxes, which makes it difficult to locate objects of multi-scale and multi-orientation. In this paper, we propose DiffSenseNet, a detection network that integrates Hierarchical Feature Fusion (HFF) and Angular Diffusion Augmentation (ADA). The HFF architecture integrates both bottom-up and top-down pathways to effectively merge features at different scales. The ADA strategy introduces directional information into noisy boxes by adding Gaussian noise to the ground truth boxes. On the DOTA and HRSC2016 datasets, DiffSenseNet achieves the accuracy of 75.89% and 86.24% mAP, respectively. Comprehensive experiments show that DiffSenseNet achieves relatively superior performance compared to previous state-of-the-art jobs.

关键词： Distributed processing Accuracy Gaussian noise Object detection Feature extraction Proposals Noise measurement Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Double-Blockchain Assisted Secure and Anonymous Data Aggregation for Fog-Enabled Smart Grid

引用

engineering 2022年第1期8卷 159-169页

作者： Siguang Chen Li Yang Chuanxin Zhao Vijayakumar Varadarajan Kun Wang Jiangsu Key Lab of Broadband Wireless Communication and Internet of Things Nanjing University of Posts and TelecommunicationsNanjing 210003China Jiangsu Engineering Research Center of Communication and Network Technology Nanjing University of Posts and TelecommunicationsNanjing 210003China Anhui Provincial Key Laboratory of Network and Information Security Anhui Normal UniversityWuhu 241000China School of Computer Science and Engineering Vellore Institute of TechnologyChennai 600127India Department of Electrical and Computer Engineering University of CaliforniaLos AngelesCA 90095USA

As a future energy system,the smart grid is designed to improve the efficiency of traditional power systems while providing more stable and reliable ***,this efficient and reliable service relies on collecting and analyzing users’electricity consumption data frequently,which induces various security and privacy *** address these challenges,we propose a double-blockchain assisted secure and anonymous data aggregation scheme for fog-enabled smart grid named ***,we design a three-tier architecture-based data aggregation framework by integrating fog computing and the blockchain,which provides strong support for achieving efficient and secure data collection in smart ***,we develop a secure and anonymous data aggregation mechanism with low computational overhead by jointly leveraging the Paillier encryption,batch aggregation signature and anonymous *** particular,the system achieves fine-grained data aggregation and provides effective support for power dispatching and price adjustment by the designed double-blockchain and two-level data ***,the superiority of the proposed scheme is illustrated by a series of security and computation cost analyses.

关键词： Blockchain Fog computing Homomorphic encryption Smart grid Anonymity

来源：评论

学校读者我要写书评

暂无评论

Split-net: Dual transformer encoder with splitting scene text image for script identification

引用

Pattern Recognition Letters 2025年 196卷 100-108页

作者： Ayush Roy Shivakumara Palaiahnakote Umapada Pal Cheng-Lin Liu Department of Computer Science and Engineering State University of New York Buffalo United States School of Science Engineering and Environment University of Salford Manchester United Kingdom Computer Vision and Pattern Recognition Indian Statistical Institute Kolkata India State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation of the Chinese Academy of Sciences Beijing China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China

Script identification is vital for understanding scenes and video images. It is challenging due to high variations in physical appearance, typeface design, complex background, distortion, and significant overlap in the characteristics of different scripts. Unlike existing models, which aim to tackle the script images utilizing the scene text image as a whole, we propose to split the image into upper and lower halves to capture the intricate differences in stroke and style of various scripts. Motivated by the accomplishments of the transformer, a modified script-style-aware Mobile-vision Transformer (M-ViT) is explored for encoding visual features of the images. To enrich the features of the transformer blocks, a novel Edge Enhanced Style Aware Channel Attention Module (EESA-CAM) has been integrated with M-ViT. Furthermore, the model fuses the features of the dual encoders (extracting features from the upper and the lower half of the images) by a dynamic weighted average procedure utilizing the gradient information of the encoders as the weights. In experiments on three standard datasets, MLe2e, CVSI2015, and SIW-13, the proposed model yielded superior performance compared to state-of-the-art models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Silicon Micro-Disk Resonator Crossbar Array for High-Speed and High-Density Photonic Convolution Processing

arXiv

引用

arXiv 2025年

作者： Huang, Long Yao, Jianping Microwave Photonics Research Laboratory School of Electrical Engineering and Computer Science University of Ottawa OttawaONK1N 6N5 Canada

Advanced artificial intelligence (AI) algorithms, particularly those based on artificial neural networks, have garnered significant attention for their potential applications in areas such as image recognition and natural language processing. Notably, neural networks make heavy use of matrix-vector multiplication (MVM) operations, causing substantial computing burden on existing electronic computing systems. Optical computing has attracted considerable attention that can perform optical-domain MVM at an ultra-high speed. In this paper, we introduce a novel silicon photonic micro-disk resonator (MDR) crossbar signal processor designed to support matrix-vector multiplication (MVM) with both high processing speed and enhanced computational density. The key innovation of the proposed MDR crossbar processor is the placement of two MDRs at each crosspoint, enabling simultaneous routing and weighting functions. This design effectively doubles the computational density, improving overall performance. We fabricate a silicon photonic MDR crossbar processor, which is employed to perform convolutional tasks in a convolutional neural network (CNN). The experimental results demonstrate that the photonic processor achieves a classification accuracy of 96% on the MNIST dataset. Additionally, it is capable of scaling to a computational speed of up to 160 tera-operations per second (TOPS) and a computational density as high as 25.6 TOPS/mm2. Our approach holds significant promise for enabling highly efficient, scalable on-chip optical computing, with broad potential applications in AI and beyond. © 2025, CC BY-NC-ND.

关键词： Silicon photonics

来源：评论

学校读者我要写书评

暂无评论

MEMS-based twisted bilayer flat optics for multidimensional optical modulation and hyperimaging

MEMS-based twisted bilayer flat optics for multidimensional ...

引用

2024 Frontiers in Optics, FiO 2024

作者： Tang, Haoning Lou, Beicheng Du, Fan Gao, Guangqi Zhang, Mingjie Ni, Xueqi Yacoby, Amir Cao, Yuan Fan, Shanhui Mazur, Eric School of Engineering and Applied Sciences Harvard University CambridgeMA02138 United States Department of Electrical Engineering and Computer Science University of California at Berkeley BerkeleyCA94720 United States National Institute for Materials Science Namiki 1-1 Ibaraki Tsukuba305-0044 Japan Department of Applied Physics Ginzton Laboratory Stanford University StanfordCA94305 United States

Our MEMS-based twistoptics device enables precise control of interlayer gaps and twist angles in photonic crystals, achieving high-accuracy, multidimensional light manipulation with significant potential in reconfigur... 详细信息

关键词： Nanophotonics

来源：评论

学校读者我要写书评

暂无评论

Differential Evolution for Optimizing Parameter Estimation in Practical D2D Channels

Differential Evolution for Optimizing Parameter Estimation i...

引用

IEEE Conference on Wireless Communications and Networking

作者： Samuel B. Ferreira Gomes Nidhi Simmons Michel Daoud Yacoub Paschalis C. Sofotasios Simon L. Cotton Wireless Technology Laboratory School of Electrical and Computer Engineering University of Campinas Campinas Brazil Centre for Wireless Innovation School of Electronics Electrical Engineering and Computer Science Queen's University Belfast Belfast UK Department of Computer and Communication Engineering C2PS and 6G Center Khalifa University Abu Dhabi UAE Department of Electrical Engineering Tampere University Tampere Finland

ISBN: (数字)9798350303582

ISBN: (纸本)9798350303599

Inspired by the theory of natural evolution, Differential Evolution (DE) functions constitute an optimization tool in the field of evolutionary algorithms. Here, we propose employing DE to estimate physically acceptable fading parameters in device-to-device (D2D) communication channels. We examine four real-world D2D propagation channel measurements obtained under various conditions: indoor, outdoor, line-of-sight (LOS), and non-LOS. Four popular fading models, $\kappa-\mu, \eta-\mu, \kappa-\mu$ /inverse gamma, and $\eta-\mu$ /inverse gamma are used to characterize these links. Two fitness functions, Kullback-Leibler divergence (KLD) and mean squared error (MSE), are utilized for evaluation. We also compare the DE with another evolutionary algorithm, namely the genetic algorithm (GA). Notably, our results demonstrate that while both algorithms deliver excellent estimation performances, DE emerges as significantly faster and more robust compared to GA. Regarding fitness performances, the algorithm, when paired with KLD, outperforms the pairing with MSE, as assessed through the minimization of the Akaike information criterion.

关键词： Fading channels Parameter estimation Metaheuristics Estimation Channel estimation Line-of-sight propagation Evolutionary computation

来源：评论

学校读者我要写书评

暂无评论

3-D Near-Field Localization by Jointly Exploiting Spatial and Temporal Information Based on a Nonuniform Cross Array

3-D Near-Field Localization by Jointly Exploiting Spatial an...

引用

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

作者： Zelong Yi Hua Chen Zhiwei Jiang Wei Liu Qing Wang Gang Wang Faculty of Electrical Engineering and Computer Science Ningbo University Ningbo China Zhejiang Key Laboratory of Mobile Network Application Technology Ningbo China School of Electronic Engineering and Computer Science Queen Mary University of London London UK School of Electronic Informatin Engineering Tianjin University Tianjin China

In this paper, an underdetermined three-dimensional (3-D) near-field source localization method is proposed, based on a two-dimensional (2-D) symmetric nonuniform cross array. Firstly, the fourth-order cumulant of the near-field observations with multiple delay lags is exploited to construct virtual far-field pseudo-observations, leading to increased degrees of freedom (DOF); then, 2-D angles of the nearfield sources are jointly estimated by employing the recently proposed sparse and parametric approach (SPA) and Vandermonde decomposition technique, eliminating the need for parameter discretization. To estimate the range term, the one-dimensional (1-D) MUSIC algorithm is applied by resorting to the conjugate symmetry property of the signal's autocorrelation function. Numerical results are provided to demonstrate the superiority of our method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Impact Assessment of NILM Methods for an Enhanced Observability of Low Voltage Distribution Networks 8

Impact Assessment of NILM Methods for an Enhanced Observabil...

引用

8th IEEE International Smart Cities Conference, ISC2 2022

作者： Moreno Jaramillo, Andres F Lopez-Lorente, Javier Brogan, Paul Laverty, David Del Rincon, Jesus Martinez Foley, Aoife School of Electronics Electrical Engineering and Computer Science Queen's University Belfast Belfast United Kingdom Pv Technology Laboratory Foss Research Centre for Sustainable Energy University of Cyprus Department of Electrical and Computer Engineering Nicosia Cyprus Applications Strategist Phasora Ltd Belfast United Kingdom School of Mechanical and Aerospace Engineering Queen's University Belfast Belfast United Kingdom

ISBN: (数字)9781665485616

ISBN: (纸本)9781665485616

The migration to low-carbon technologies in secondary (low-voltage) distribution networks relies on high integration of distributed energy resources (DER), including distributed generation and low-carbon loads. The electrification of power and transport sectors has led to an increased deployment of solar photovoltaic systems and electric vehicles causing unconventional dynamics in distribution networks. Thus, distribution network operators are facing new technical challenges such as reverse power flow, higher load peaks, and power quality issues. This has attracted research interest to investigate the implementation of innovative techniques such as non-intrusive load monitoring (NILM) methods. These techniques contribute to an enhanced and scalable monitoring of DER in electrical networks. Current literature based NILM models have achieved significantly high performance for both electromobility and photovoltaic systems, which makes them eligible for deployment from a technical standpoint. However, a critical analysis of the considerations for a feasible implementation of these methodologies has not been covered. Thus, this work studies the techno-economic implications for the large-scale implementation of NILM methods at the customer level and low voltage side of distribution transformers. While a substantial investment is associated with a better monitored low voltage distribution network, there are multiple benefits related to technical, economic, social, and environmental aspects. © 2022 IEEE.

关键词： Electric power transmission networks

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：