检索结果-内蒙古大学图书馆

SSRN 2023年

作者： Li, Hang Li, Wentong Lv, Yina Liu, Jialin Yang, Long Shi, Liang East China Normal University Shanghai200062 China Software/Hardware Co-design Engineering Research Center Ministry of Education The School of Computer Science and Technology East China Normal University Shanghai China

The mobile distributed file system allows users to share files among multiple mobile devices through simple read/write operations. However, it requires the connected remote devices to be online all the time to ensure timely file browsing, which incurs significant energy consumption. This is unacceptable for battery-powered mobile devices. To address this issue, we propose ESPCache, an energy-efficient stateful persistent cache framework for mobile distributed file systems. It consists of four key techniques. First, a proactive cache invalidation mechanism is designed to ensure optimistic access to the persistent cache locally, which greatly reduces non-essential file requests. Second, a lazy cache synchronization policy is designed to reorganize writeback requests, which ensures that remote devices remain in a low-power state for a longer time. Finally, a cache admission and eviction scheme is proposed, which considers both file access frequency and recency, and an adaptable file prefetching scheme is designed to quickly recover invalidated cache files. Evaluations on real devices show that ESPCache maintains at least 60% of sleep time for the remote device and greatly extends the interval between two wake-ups, regardless of the frequency of file accesses. Compared with the state-of-the-art, the energy consumption of remote devices can be reduced by 33.6%. © 2023, The Authors. All rights reserved.

关键词： Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

Permutation Invariant Training for Paraphrase Identification

Permutation Invariant Training for Paraphrase Identification

引用

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

作者： Jun Bai Chuantao Yin Hanhua Hong Jianfei Zhang Chen Li Yanmeng Wang Wenge Rong State Key Laboratory of Software Development Environment Beihang University Beijing China School of Computer Science and Engineering Beihang University Beijing China Sino-French Engineer School Beihang University Beijing China Ping An Technology Beijing China

Identifying sentences sharing similar meanings is crucial to speech and text understandings. Although currently popular cross-encoder solutions with pre-trained language models as backbone have achieved remarkable performance, they suffer from the lack of the permutation invariance or symmetry that is one of the most important inductive biases to such task. To alleviate this issue, in this research we propose a permutation invariant training framework, in which a symmetry regularization is introduced during training that forces the model to produce the same predictions for input sentence pairs in both forward and backward directions. Empirical studies exhibit improved performance over competitive baselines.

关键词： Training Signal processing Predictive models Acoustics Task analysis Speech processing

来源：评论

学校读者我要写书评

暂无评论

SPA: A Graph Spectral Alignment Perspective for Domain Adaptation

arXiv

引用

arXiv 2023年

作者： Xiao, Zhiqing Wang, Haobo Jin, Ying Feng, Lei Chen, Gang Huang, Fei Zhao, Junbo College of Computer Science and Technology Zhejiang University China School of Software Technology Zhejiang University China Key Lab of Intelligent Computing based Big Data of Zhejiang Province Zhejiang University China CUHK-SenseTime Joint Lab The Chinese University of Hong Kong Hong Kong School of Computer Science and Engineering Nanyang Technological University Singapore Alibaba Group China

Unsupervised domain adaptation (UDA) is a pivotal form in machine learning to extend the in-domain model to the distinctive target domains where the data distributions differ. Most prior works focus on capturing the inter-domain transferability but largely overlook rich intra-domain structures, which empirically results in even worse discriminability. In this work, we introduce a novel graph SPectral Alignment (SPA) framework to tackle the tradeoff. The core of our method is briefly condensed as follows: (i)-by casting the DA problem to graph primitives, SPA composes a coarse graph alignment mechanism with a novel spectral regularizer towards aligning the domain graphs in eigenspaces;(ii)-we further develop a fine-grained message propagation module — upon a novel neighbor-aware self-training mechanism — in order for enhanced discriminability in the target domain. On standardized benchmarks, the extensive experiments of SPA demonstrate that its performance has surpassed the existing cutting-edge DA methods. Coupled with dense model analysis, we conclude that our approach indeed possesses superior efficacy, robustness, discriminability, and transferability. Code and data are available at: https://***/CrownX/SPA. 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

Spatial-temporal Consistency based Crowd Emotion Recognition

Spatial-temporal Consistency based Crowd Emotion Recognition

引用

International Conference on computer Supported Cooperative Work in Design

作者： Chuanmiao Zhao Guijuan Zhang Lei Lv Hong Liu Dianjie Lu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

In crowd evacuation scenarios, it is an effective way to ensure evacuation safety by recognizing real crowd emotions and then taking measures such as emotional infection to reduce crowd panic. However, factors such as distance, exposure, angle, and occlusion can lead to incomplete face expression information collection, thus failing to accurately identify crowd emotions. Therefore, it is still a very challenging problem to recognize individual emotions in the case of incomplete face expression information collection, and thus accurately identify crowd emotions during evacuation. To solve this problem, we propose a spatial-temporal consistency-based crowd emotion recognition method to accurately identify the real emotions of the crowd. First, for video frames that can capture the complete facial expression information, we use the residual network to accurately identify the individual emotion values in each frame. For video frames that cannot capture the complete facial expression information, we propose an individual emotion calculation model based on spatial-temporal consistency to calculate the individual emotion values in each frame. Second, we define the crowd panic level and obtain the real crowd emotion by calculation. Finally, we implement an end-to-end crowd panic emotion recognition system to verify our method. The experimental results show that the method can accurately calculate the crowd panic level, which is important for guiding crowd evacuation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-scale Two-Way Deblurring Network for Non-uniform Single Image Deblurring 14th

Multi-scale Two-Way Deblurring Network for Non-uniform Singl...

引用

14th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2021

作者： Cheng, Zhongzhe Luo, Bing Xu, Li Li, Siwei Xiao, Kunshu Pei, Zheng The Center for Radio Administration Technology Development School of Computer and Software Engineering Xihua University Chengdu610039 China The School of Science Xihua University Chengdu610039 China Xihua College of Xihua University Chengdu610039 China

ISBN: (纸本)9783030898137

We propose a new and effective image deblurring network based on deep learning. The motivation of this work is based on traditional algorithms and deep learning which take an easy-to-difficult approach to image deblurring. In traditional algorithms, a rough blur kernel is obtained first, and then a precise blur kernel is gradually refined. In deep learning, the pyramid structure is adopted to restore clear images from easy to difficult. We hope to recover the clear image by two-way approximation. One network recovers the roughly clear image from the blurred image, and the other network recovers part of the structural information from the blank image, and finally the two networks are added together to obtain the clear image. Experiments show that since we decomposed the original deblurring task into two different tasks, the network performance has been effectively improved. Compared with other latest networks, our network can get clearer images. © 2021, ICST Institute for computer sciences, Social Informatics and Telecommunications Engineering.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalization

arXiv

引用

arXiv 2024年

作者： Li, Jiahe Zhang, Jiawei Bai, Xiao Zheng, Jin Ning, Xin Zhou, Jun Gu, Lin School of Computer Science and Engineering State Key Laboratory of Complex & Critical Software Environment Jiangxi Research Institute Beihang University China Institute of Semiconductors Chinese Academy of Sciences China School of Information and Communication Technology Griffith University Australia RIKEN AIP Japan The University of Tokyo Japan

Radiance fields have demonstrated impressive performance in synthesizing novel views from sparse input views, yet prevailing methods suffer from high training costs and slow inference speed. This paper introduces DNGaussian, a depth-regularized framework based on 3D Gaussian radiance fields, offering real-time and high-quality few-shot novel view synthesis at low costs. Our motivation stems from the highly efficient representation and surprising quality of the recent 3D Gaussian Splatting, despite it will encounter a geometry degradation when input views decrease. In the Gaussian radiance fields, we find this degradation in scene geometry primarily lined to the positioning of Gaussian primitives and can be mitigated by depth constraint. Consequently, we propose a Hard and Soft Depth Regularization to restore accurate scene geometry under coarse monocular depth supervision while maintaining a fine-grained color appearance. To further refine detailed geometry reshaping, we introduce Global-Local Depth Normalization, enhancing the focus on small local depth changes. Extensive experiments on LLFF, DTU, and Blender datasets demonstrate that DNGaussian outperforms state-of-the-art methods, achieving comparable or better results with significantly reduced memory cost, a 25× reduction in training time, and over 3000× faster rendering speed. Code is available at: https://***/Fictionarry/DNGaussian. Copyright © 2024, The Authors. All rights reserved.

关键词： Geometry

来源：评论

学校读者我要写书评

暂无评论

Singular Value Decomposition Based Pedestrian Trajectory Prediction

Singular Value Decomposition Based Pedestrian Trajectory Pre...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Heng Wang Guijuan Zhang Hong Liu Dianjie Lu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

Numerous cameras deployed in venues can collect video data, which can be analyzed to help evacuate people in emergency situations. Most surveillance videos can provide data support for pedestrian trajectory prediction. However, data-driven prediction method do not consider the impact of personalization on pedestrian trajectories, which results in ignoring individual personalization. How to accurately analyze the differences among pedestrians and yet accurately predict pedestrian trajectories is a challenging problem. To solve the above problems, we propose an singular value decomposition (SVD) based pedestrian trajectory prediction method, which introduces matrix decomposition to pedestrian trajectory prediction for the first time. Thus, This method analyzes the impact of pedestrian personalization on motion trajectories by mining the interaction patterns between pedestrians and environment. Firstly, we propose an information collection method based on environmental semantics, which can extract scene information from historical video data to construct an environmental information matrix. Secondly, we propose an SVD-based individual environmental feature preference method, which uses singular value decomposition methods to mine the data and analyze personalized pedestrian motion patterns to construct a personalized individual preference matrix. Finally, we built an personalized trajectory prediction method to predict pedestrian movement trajectories. The experimental results show that the method can not only analyze the effect of personalization on pedestrian movement but also accurately predict the pedestrian movement trajectory.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Federated graph learning for cross-domain recommendation 24

Federated graph learning for cross-domain recommendation

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Ziqi Yang Zhaopeng Peng Zihui Wang Jianzhong Qi Chaochao Chen Weike Pan Chenglu Wen Cheng Wang Xiaoliang Fan Fujian Key Laboratory of Sensing and Computing for Smart Cities Xiamen University China and Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of China Xiamen University China School of Computing and Information Systems The University of Melbourne Australia College of Computer Science and Technology Zhejiang University Hangzhou China College of Computer Science and Software Engineering Shenzhen University Shenzhen China

ISBN: (纸本)9798331314385

Cross-domain recommendation (CDR) offers a promising solution to the data sparsity problem by enabling knowledge transfer between source and target domains. However, many recent CDR models overlook crucial issues such as privacy as well as the risk of negative transfer (which negatively impact model performance), especially in multi-domain settings. To address these challenges, we propose FedGCDR, a novel federated graph learning framework that securely and effectively leverages positive knowledge from multiple source domains. First, we design a positive knowledge transfer module that ensures privacy during inter-domain knowledge transmission. This module employs differential privacy-based knowledge extraction combined with a feature mapping mechanism, transforming source domain embeddings from federated graph attention networks into reliable domain knowledge. Second, we design a knowledge activation module to filter out potential harmful or conflicting knowledge from source domains, addressing the issues of negative transfer. This module enhances target domain training by expanding the graph of the target domain to generate reliable domain attentions and fine-tunes the target model for improved negative knowledge filtering and more accurate predictions. We conduct extensive experiments on 16 popular domains of the Amazon dataset, demonstrating that FedGCDR significantly outperforms state-of-the-art methods. We open source the code at https://***/LafinHana/FedGCDR.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Complex Belief Jensen-Shannon Divergence in Complex Evidence Theory for Decision-Making

A Complex Belief Jensen-Shannon Divergence in Complex Eviden...

引用

IEEE International Conference on Unmanned Systems (ICUS)

作者： Linlu Gao Mingrui Yin Fuyuan Xiao Zehong Cao The School of Big Data and Software Engineering Chongqing University Chongqing China College of Computer Science and Technology Zhejiang University Hangzhou China STEM Mawson Lakes Campus University of South Australia Adelaide Australia

Effective target detection and recognition are essential in remote sensing image field, and complex evidence theory (CET) is widely used for this purpose. However, measuring conflict between complex basic belief assignments (CBBAs) in CET is challenging. This study proposes a complex belief Jensen-Shannon divergence based on the complex Pignistic transformation to measure conflict, accounting for quantum interference effects in CBBAs. We analyze the properties of the CBJS divergence, including boundedness, nondegeneracy, and symmetry. Numerical examples are presented to show the effectiveness of the conflict measure method and the potential of improving the robustness of intelligent interpretation system. Moreover, an algorithm for decision-making is presented and applied in pattern recognition illustrating its superiority.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Reality3DSketch: Rapid 3D Modeling of Objects from Single Freehand Sketches

arXiv

引用

arXiv 2023年

作者： Chen, Tianrun Ding, Chaotao Zhu, Lanyun Zang, Ying Liao, Yiyi Li, Zejian Sun, Lingyun Technology Co. LTD 310027 China The School of Software Technology Zhejiang University 310027 China The School of Information Engineering Huzhou University 313000 China The College of Information Science and Electronic Engineering Zhejiang University 310027 China The Information Systems Technology and Design Pillar Singapore University of Technology and Design Singapore487372 Singapore The College of Computer Science and Technology Zhejiang University 310027 China

The emerging trend of AR/VR places great demands on 3D content. However, most existing software requires expertise and is difficult for novice users to use. In this paper, we aim to create sketch-based modeling tools for user-friendly 3D modeling. We introduce Reality3DSketch with a novel application of an immersive 3D modeling experience, in which a user can capture the surrounding scene using a monocular RGB camera and can draw a single sketch of an object in the real-time reconstructed 3D scene. A 3D object is generated and placed in the desired location, enabled by our novel neural network with the input of a single sketch. Our neural network can predict the pose of a drawing and can turn a single sketch into a 3D model with view and structural awareness, which addresses the challenge of sparse sketch input and view ambiguity. We conducted extensive experiments synthetic and real-world datasets and achieved state-of-the-art (SOTA) results in both sketch view estimation and 3D modeling performance. According to our user study, our method of performing 3D modeling in a scene is >5x faster than conventional methods. Users are also more satisfied with the generated 3D model than the results of existing methods. © 2023, CC BY-NC-ND.

关键词： 3D modeling

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：