检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Liu, Niqi Liu, Fang Ji, Wenqi Du, Xinxin Liu, Xu Zhao, Guozhen Mu, Wenting Liu, Yong-Jin BNRist Department of Computer Science and Technology MOE-Key Laboratory of Pervasive Computing Tsinghua University China State Key Laboratory of Media Convergence and Communication Communication University of China China Multimodal Sensing and Computing Laboratory Beijing China CAS Key Laboratory of Behavioral Science Institute of Psychology China Department of Psychology Tsinghua University China

The robust generalization of deep learning models in the presence of inherent noise remains a significant challenge, especially when labels are subjective and noise is indiscernible in natural settings. This problem is particularly pronounced in many practical applications. In this paper, we address a special and important scenario of monitoring suicidal ideation, where time-series data, such as photoplethysmography (PPG), is susceptible to such noise. Current methods predominantly focus on image and text data or address artificially introduced noise, neglecting the complexities of natural noise in time-series analysis. To tackle this, we introduce a novel neural network model tailored for analyzing noisy physiological time-series data, named TNANet, which merges advanced encoding techniques with confidence learning, enhancing prediction accuracy. Another contribution of our work is the collection of a specialized dataset of PPG signals derived from real-world environments for suicidal ideation prediction. Employing this dataset, our TNANet achieves the prediction accuracy of 63.33% in a binary classification task, outperforming state-of-the-art models. Furthermore, comprehensive evaluations were conducted on three other well-known public datasets with artificially introduced noise to rigorously test the TNANet's capabilities. These tests consistently demonstrated TNANet's superior performance by achieving an accuracy improvement of more than 10% compared to baseline methods. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Research on Error Handling Techniques for Speech Interaction

Research on Error Handling Techniques for Speech Interaction

引用

computer Systems (ICCS), IEEE International Conference on

作者： Rui Ma Xuechao Zou Pin Tao Yixiang Tu Department of Computer Science and Technology Tsinghua University Beijing China Department of Computer Technology and Applications Qinghai University Xining China Key Laboratory of Pervasive Computing Ministry of Education Beijing China

With the rapid advancement of speech recognition and semantic understanding, speech interaction systems have gained widespread usage. This paper investigates the use of intelligent error handling approaches in speech interaction systems when faced with recognition or comprehension challenges. A case study focusing on the common task of “electronic spreadsheet editing” is conducted, wherein an experimental environment is developed by combining an open-source neural network code repository with a third-party speech recognition interface. Drawing from historical experience, a “triplet” paradigm is constructed to parse user command texts, and an improved approach for intelligent selection of error recovery strategies is proposed. This approach automatically determines the appropriate strategy for different speech interaction errors based on user logs. Experiments on 20 users demonstrate that our proposed intelligent error handling techniques for speech interaction, exhibit performance similar to empirically designed systems in practical tasks. Moreover, they significantly enhance user satisfaction and task completion efficiency when utilizing the system.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DSD-MatchingNet:Deformable sparse-to-dense feature matching for learning accurate correspondences

引用

Virtual Reality & Intelligent Hardware 2022年第5期4卷 432-443页

作者： Yicheng ZHAO Han ZHANG Ping LU Ping LI Enhua WU Bin SHENG Department of Computer Science and Engineering Shanghai Jiao Tong UniversityShanghai 200240China ZTE Corporation Shenzhen 518057China State Key Laboratory of Mobile Network and Mobile Multimedia Technology Shenzhen 518057China Department of Computing The Hong Kong Polytechnic UniversityHong Kong 999077China School of Design The Hong Kong Polytechnic UniversityHong Kong 999077China State Key Laboratory of Computer Science Institute of SoftwareChinese Academy of SciencesBeijing 100190China Faculty of Science and Technology University of MacaoMacao 999078China

Background Exploring correspondences across multiview images is the basis of various computer vision ***,most existing methods have limited accuracy under challenging *** To learn more robust and accurate correspondences,we propose DSD-MatchingNet for local feature matching in this ***,we develop a deformable feature extraction module to obtain multilevel feature maps,which harvest contextual information from dynamic receptive *** dynamic receptive fields provided by the deformable convolution network ensure that our method obtains dense and robust ***,we utilize sparse-to-dense matching with symmetry of correspondence to implement accurate pixel-level matching,which enables our method to produce more accurate *** Experiments show that our proposed DSD-MatchingNet achieves a better performance on the image matching benchmark,as well as on the visual localization ***,our method achieved 91.3%mean matching accuracy on the HPatches dataset and 99.3%visual localization recalls on the Aachen Day-Night dataset.

关键词： Image matching Deformable convolution network Sparse-to-dense matching

来源：评论

学校读者我要写书评

暂无评论

KNN-Based Collaborative Filtering for Fine-Grained Intelligent Grad-School Recommendation System 19th

KNN-Based Collaborative Filtering for Fine-Grained Intellig...

引用

19th EAI International Conference on Heterogeneous Networking for Quality, Reliability, Security and Robustness, QShine 2023

作者： Xu, Jinfeng Liu, Jiyi Ma, Zixiao Wang, Yuyang Wang, Wei Ngai, Edith Department of Electrical and Electronic Engineering The University of Hong Kong 999077 Hong Kong School of Software Engineering Sun Yat-Sen University Guangdong Zhuhai528406 China Department of Electrical and Computer Engineering Duke University DurhamNC27708 United States Department of Computer Science Cornell University IthacaNY14853 United States Guangdong-Hong Kong-Macao Joint Laboratory for Emotional Intelligence and Pervasive Computing Artificial Intelligence Research Institute Shenzhen MSU-BITUniversity Guangdong Shenzhen518172 China

ISBN: (纸本)9783031651250

The development of the Internet has led to information overload, and how to filter and sift information is a rigorous requirement in all fields. In response to this challenge, recommendation systems have emerged as a valuable tool, offering personalized content and services by efficiently searching and processing dynamically generated information. For students applying to grad schools, finding relevant information can be time-consuming and unreliable from official websites or forums. In light of these challenges, we present a novel solution in the form of an application recommendation platform. Our proposed platform leverages specific open-source datasets and real-time information from platform users using KNN (K-Nearest Neighbor) and CF (Collaborative Filtering) techniques to provide recommendations based on users’ individual backgrounds, we aim to reduce the complexity inherent in information retrieval while simultaneously enhancing the relevance of the recommendations delivered to users. Specifically, we first collect user behavior data, then we will construct the data model and perform some preprocessing on it. Calculate the user similarity, and find out the K-nearest neighbors and rate based on K-nearest neighbors, finally, the recommendation engine is used to calculate the highest-rated items to be recommended to the users. © ICST Institute for computer sciences, Social Informatics and Telecommunications Engineering 2024.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

THUD++: Large-Scale Dynamic Indoor Scene Dataset and Benchmark for mobile Robots

arXiv

引用

arXiv 2024年

作者： Li, Zeshun Li, Fuhao Zhang, Wanting Zheng, Zijie Liu, Xueping Liu, Yongjin Zeng, Long Department of Advanced Manufacturing Shenzhen International Graduate School Tsinghua University Shenzhen518000 China BNRist MOE Key Laboratory of Pervasive Computing Department of Computer Science and Technology Tsinghua University Beijing100084 China

Most existing mobile robotic datasets primarily capture static scenes, limiting their utility for evaluating robotic performance in dynamic environments. To address this, we present a mobile robot oriented large-scale indoor dataset, denoted as THUD++ (TsingHua University Dynamic) robotic dataset, for dynamic scene understanding. Our current dataset includes 13 large-scale dynamic scenarios, combining both real-world and synthetic data collected with a real robot platform and a physical simulation platform, respectively. The RGB-D dataset comprises over 90K image frames, 20M 2D/3D bounding boxes of static and dynamic objects, camera poses, and IMU. The trajectory dataset covers over 6,000 pedestrian trajectories in indoor scenes. Additionally, the dataset is augmented with a Unity3D-based simulation platform, allowing researchers to create custom scenes and test algorithms in a controlled environment. We evaluate state-of-the-art methods on THUD++ across mainstream indoor scene understanding tasks, e.g., 3D object detection, semantic segmentation, relocalization, pedestrian trajectory prediction, and navigation. Our experiments highlight the challenges mobile robots encounter in indoor environments, especially when navigating in complex, crowded, and dynamic scenes. By sharing this dataset, we aim to accelerate the development and testing of mobile robot algorithms, contributing to real-world robotic applications. The dataset is shared at: https://***/***/. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Enhancing Multilabel ECG Classification via Task-Guided Lead Correlations in Internet of Medical Things

引用

IEEE Internet of Things Journal 2025年第12期12卷 20544-20555页

作者： Yuan, Xiaoyan Wang, Wei Chen, Junxin Fang, Kai Kashif Bashir, Ali Mondal, Tapas Hu, Xiping Jamal Deen, M. Beijing Institute of Technology School of Medical Technology Beijing100081 China Shenzhen MSU-BIT University Guangdong-Hong Kong-Macao Joint Laboratory for Emotion Intelligence and Pervasive Computing Artificial Intelligence Research Institute Guangdong Shenzhen518172 China Dalian University of Technology School of Software Dalian116621 China Zhejiang A&F University College of Mathematics and Computer Science Hangzhou311300 China Manchester Metropolitan University Department of Computing and Mathematics ManchesterM15 6BH United Kingdom Chitkara University Institute of Engineering and Technology Chitkara University Centre for Research Impact and Outcome Rajpura140401 India McMaster University Faculty of Health Science Department of Pediatrics ON HamiltonL8S 4L8 Canada McMaster University McMaster School of Biomedical Engineering Electrical and Computer Engineering Department ON HamiltonL8S 4L8 Canada

With the rise of the Internet of Things, wearable devices have enabled real-time health monitoring, particularly through physiological signals like electrocardiograms (ECG). The standard 12-lead ECG records the electrical activity of the heart from multiple perspectives, providing valuable insights into cardiac health. However, existing 12-lead ECG analysis methods often treat leads as channel-level arrangements or rely on spatial adjacency to predefine lead connections, limiting their ability to capture the complex spatial and functional relationships between leads fully. To address this limitation, we propose TGLLNet, a task-driven model that automatically learns interlead relationships to improve multilabel ECG classification. TGLLNet adaptively learns lead connectivity patterns and relational strengths, enhancing ECG representation and improving model generalizability across tasks. Specifically, TGLLNet employs a temporal graph construction module to convert ecg signals into temporal graphs and uses a residual pyramid graph convolution module for multilevel graph embeddings, utilizing a graph convolutional network with independently learnable adjacency matrices. Combined with a temporal context convolution module, TGLLNet captures spatio-temporal dependencies, significantly improving ECG representation. Experimental results on seven tasks from PTB-XL and CPSC2018 datasets demonstrate that TGLLNet outperforms existing methods, showing superior generalizability across different tasks. Our code is available at https://***/rosemary333/TGLLnet. © 2014 IEEE.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

Subjective Quality of Experience Assessment in mobile Cloud Games

Subjective Quality of Experience Assessment in Mobile Cloud ...

引用

GLOBECOM 2022 - 2022 IEEE Global Communications Conference

作者： Henrique Souza Rossi Niclas Ögren Karan Mitra Irina Cotanis Christer Åhlund Per Johansson Department of Computer Science Electrical and Space Engineering Mobile and Pervasive Computing Luleå University of Technology Skellefteå Sweden Infovista Skellefteå Sweden

ISBN: (纸本)9781665435413

The rise of mobile cloud gaming (MCG) has necessitated understanding its impact on mobile network design and deployment for end users' QoE maximization. MCG is a dynamic service that requires stringent quality from network operators. Therefore, this paper investigates the subjective QoE of MCG over mobile networks played on smartphones. We conducted subjective tests (N=31); our results indicate that MCG is affected differently by QoS attributes such as packet loss (PL), round trip time (RTT) and jitter compared to cloud games and online mobile games. We identify that RTT values above 100 milliseconds significantly impact users' QoE, measured via the mean opinion score (MOS). Further, lower RTT values with high PL; and higher RTT values with low PL cause a strong negative effect on MOS. Lastly, bursty jitter seems to affect the MOS, while random jitter does not significantly impact MOS.

关键词： Packet loss Quality of service Jitter Cloud gaming Quality of experience Global communication Smart phones

来源：评论

学校读者我要写书评

暂无评论

A Data Contribution-Based Adaptive Federated Learning Approach for Wearable Activity Recognition

A Data Contribution-Based Adaptive Federated Learning Approa...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Yi Zhao Chunyu Hu Xiaodong Yang Lin Yuan Xiang Tian Tianlei Gao Yiqiang Chen Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center (National Supercomputer Center in Jinan) Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China Beijing Key Laboratory of Mobile Computing and Pervasive Device Institute of Computing Technology Chinese Academy of Sciences Beijing China Shandong Artificial Intelligence Institute Qilu University of Technology (Shandong Academy of Sciences) Jinan China College of Computer Science and Engineering Shandong University of Science and Technology Qingdao China

ISBN: (数字)9798331513054

ISBN: (纸本)9798331513061

Wearable activity recognition is crucial for ubiquitous computing, enhancing human-machine interaction, medical monitoring, and personalized services. As wearable devices collect user activity data that often contain personal privacy information, federated learning (FL) is increasingly applied to protect user data privacy. However, in real-world scenarios, users' data are commonly exhibit heterogeneity, manifesting as non-independent and identically distributed (non-IID) characteristics, which presents challenges for FL methods. Traditional FL client selection approaches with heterogeneous data can cause global model drift, reducing the accuracy of activity recognition models. In this paper, we propose Data Contribution-Based Federated Learning (DCBFL) method, an adaptive FL training approach by selecting clients to counter the problem caused by heterogeneous data. Specifically, we first utilize a conditional generator on the server to construct an auxiliary dataset, which is used to train an auxiliary model as a benchmark to measure the degree of heterogeneity in each client's data. Furthermore, we reasonably differentiate the data contributions of clients based on the degree of data heterogeneity and select suitable clients for FL training, effectively utilizing heterogeneous data information, mitigating global model drift. The comprehensive experiments are conducted on five public activity recognition datasets under non-IID conditions in this work. The experimental results show that DCBFL outperforms existing baseline methods, showcasing superior performance.

关键词： Training Adaptation models Data privacy Accuracy Federated learning Activity recognition Ubiquitous computing Data models Biomedical monitoring Wearable devices

来源：评论

学校读者我要写书评

暂无评论

Interpersonal Relationship Analysis with Dyadic EEG Signals via Learning Spatial-Temporal Patterns

arXiv

引用

arXiv 2024年

作者： Ji, Wenqi Liu, Fang Du, Xinxin Liu, Niqi Zhou, Chao Yu, Minjing Zhao, Guozhen Liu, Yong-Jin BNRist Department of Computer Science and Technology MOE Key Laboratory of Pervasive Computing Tsinghua University China State Key Laboratory of Media Convergence and Communication Communication University of China China Institute of Software Chinese Academy of Sciences China College of Intelligence and Computing Tianjin University China CAS Key Laboratory of Behavioral Science Institute of Psychology China

Interpersonal relationship quality is pivotal in social and occupational contexts. Existing analysis of interpersonal relationships mostly rely on subjective self-reports, whereas objective quantification remains challenging. In this paper, we propose a novel social relationship analysis framework using spatio-temporal patterns derived from dyadic EEG signals, which can be applied to quantitatively measure team cooperation in corporate team building, and evaluate interpersonal dynamics between therapists and patients in psychiatric therapy. First, we constructed a dyadic-EEG dataset from 72 pairs of participants with two relationships (stranger or friend) when watching emotional videos simultaneously. Then we proposed a deep neural network on dyadic-subject EEG signals, in which we combine the dynamic graph convolutional neural network for characterizing the interpersonal relationships among the EEG channels and 1-dimension convolution for extracting the information from the time sequence. To obtain the feature vectors from two EEG recordings that well represent the relationship of two subjects, we integrate deep canonical correlation analysis and triplet loss for training the network. Experimental results show that the social relationship type (stranger or friend) between two individuals can be effectively identified through their EEG data. Copyright © 2024, The Authors. All rights reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

VoxInstruct: Expressive Human Instruction-to-Speech Generation with Unified Multilingual Codec Language Modelling

arXiv

引用

arXiv 2024年

作者： Zhou, Yixuan Qin, Xiaoyu Jin, Zeyu Zhou, Shuoyi Lei, Shun Zhou, Songtao Wu, Zhiyong Jia, Jia Shenzhen International Graduate School Tsinghua University Shenzhen China Department of Computer Science and Technology Tsinghua University Beijing China BNRist Tsinghua University Key Laboratory of Pervasive Computing Ministry of Education Beijing China

Recent AIGC systems possess the capability to generate digital multimedia content based on human language instructions, such as text, image and video. However, when it comes to speech, existing methods related to human instruction-to-speech generation exhibit two limitations. Firstly, they require the division of inputs into content prompt (transcript) and description prompt (style and speaker), instead of directly supporting human instruction. This division is less natural in form and does not align with other AIGC models. Secondly, the practice of utilizing an independent description prompt to model speech style, without considering the transcript content, restricts the ability to control speech at a fine-grained level. To address these limitations, we propose VoxInstruct, a novel unified multilingual codec language modeling framework that extends traditional text-to-speech tasks into a general human instruction-to-speech task. Our approach enhances the expressiveness of human instruction-guided speech generation and aligns the speech generation paradigm with other modalities. To enable the model to automatically extract the content of synthesized speech from raw text instructions, we introduce speech semantic tokens as an intermediate representation for instruction-to-content guidance. We also incorporate multiple Classifier-Free Guidance (CFG) strategies into our codec language model, which strengthens the generated speech following human instructions. Furthermore, our model architecture and training strategies allow for the simultaneous support of combining speech prompt and descriptive human instruction for expressive speech synthesis, which is a first-of-its-kind attempt. Codes, models and demos are at:https://***/thuhcsi/VoxInstruct. © 2024, CC BY.

关键词： C (programming language)

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：