检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Wang, Yaxian Ding, Henghui He, Shuting Jiang, Xudong Wei, Bifan Liu, Jun School of Computer Science and Technology Xi'an Jiaotong University China Ministry of Education Key Laboratory of Intelligent Networks and Network Security Xi'an Jiaotong University China Institute of Big Data Fudan University China Shanghai University of Finance and Economics China Nanyang Technological University Singapore School of Continuing Education Xi'an Jiaotong University China Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Xi'an Jiaotong University China

ISBN: (纸本)157735897X

In this work, we address the challenging task of Generalized Referring Expression Comprehension (GREC). Compared to the classic Referring Expression Comprehension (REC) that focuses on single-target expressions, GREC extends the scope to a more practical setting by further encompassing no-target and multi-target expressions. Existing REC methods face challenges in handling the complex cases encountered in GREC, primarily due to their fixed output and limitations in multi-modal representations. To address these issues, we propose a Hierarchical Alignment-enhanced Adaptive Grounding Network (HieA2G) for GREC, which can flexibly deal with various types of referring expressions. First, a Hierarchical Multi-modal Semantic Alignment (HMSA) module is proposed to incorporate three levels of alignments, including word-object, phrase-object, and text-image alignment. It enables hierarchical cross-modal interactions across multiple levels to achieve comprehensive and robust multi-modal understanding, greatly enhancing grounding ability for complex cases. Then, to address the varying number of target objects in GREC, we introduce an Adaptive Grounding Counter (AGC) to dynamically determine the number of output targets. Additionally, an auxiliary contrastive loss is employed in AGC to enhance object-counting ability by pulling in multi-modal features with the same counting and pushing away those with different counting. Extensive experimental results show that HieA2G achieves new state-of-the-art performance on the challenging GREC task and also the other 4 tasks, including REC, Phrase Grounding, Referring Expression Segmentation (RES), and Generalized Referring Expression Segmentation (GRES), demonstrating the remarkable superiority and generalizability of the proposed HieA2G. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Modal analysis

来源：评论

学校读者我要写书评

暂无评论

KNN-MMD: Cross Domain Wireless Sensing via Local Distribution Alignment

arXiv

引用

arXiv 2024年

作者： Zhao, Zijian Cai, Zhijie Chen, Tingwei Li, Xiaoyang Li, Hang Chen, Qimei Zhu, Guangxu Shenzhen Research Institute of Big Data Shenzhen518115 China School of Computer Science and Engineering Sun Yat-sen University Guangzhou510275 China Shenzhen518115 China School of Electronic Information Wuhan University Wuhan430072 China

Wireless sensing has recently found widespread applications in diverse environments, including homes, offices, and public spaces. By analyzing patterns in channel state information (CSI), it is possible to infer human actions for tasks such as person identification, gesture recognition, and fall detection. However, CSI is highly sensitive to environmental changes, where even minor alterations can significantly distort the CSI patterns. This sensitivity often leads to performance degradation or outright failure when applying wireless sensing models trained in one environment to another. To address this challenge, Domain Alignment (DAL) has been widely adopted for cross-domain classification tasks, as it focuses on aligning the global distributions of the source and target domains in feature space. Despite its popularity, DAL often neglects inter-category relationships, which can lead to misalignment between categories across domains, even when global alignment is achieved. To overcome these limitations, we propose K-Nearest Neighbors Maximum Mean Discrepancy (KNN-MMD), a novel few-shot method for cross-domain wireless sensing. Our approach begins by constructing a "help set" using K-Nearest Neighbors (KNN) from the target domain, enabling local alignment between the source and target domains within each category using Maximum Mean Discrepancy (MMD). Additionally, we address a key instability issue commonly observed in cross-domain methods, where model performance fluctuates sharply between epochs. Further, most existing methods struggle to determine an optimal stopping point during training due to the absence of labeled data from the target domain. Our method resolves this by excluding the support set from the target domain during training and employing it as a validation set to determine the stopping criterion. We evaluate the effectiveness of the proposed method across several cross-domain Wi-Fi sensing tasks, including gesture recognition, person identification, f

关键词： Gesture recognition

来源：评论

学校读者我要写书评

暂无评论

MMFNet: A Multi-modal and Multi-attention Fusion Network for Multi-task Skin Disease Classification

MMFNet: A Multi-modal and Multi-attention Fusion Network for...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Xinlei Zhu Pengjiang Qian Chuang Wang Li Liu Ming Gao Weiwei Cai Eddie Yin-Kwee Ng School of Artificial Intelligence and Computer Science Jiangnan University Wuxi China Engineering Research Center of Intelligent Technology for Healthcare Ministry of Education Jiangnan University Wuxi China Big Data Center Affiliated Hospital of Jiangnan University Wuxi China School of Mechanical and Aerospace Engineering Nanyang Technological University Nanyang Singapore

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Skin diseases rank among the most prevalent ailments in humans, which underscores the critical importance of early detection and diagnosis. Considering clinical practice, the integration of information from various modalities holds significant potential to enhance diagnostic accuracy and precision. In this study, we propose a multi-modal and multi-attention fusion network for multi-task skin disease classification. We initially devise the Shifted Window Cross-attention Fusion (SWCF) module based on the shifted window self-attention mechanism, leveraging the advantages of the attention mechanism to learn the correlations between different modalities and fully integrate multi-modal features. Subsequently, taking into account the heterogeneity of the data, we employ the Heterogeneous data Cross-attention Fusion (HDCF) module using the idea of information decoupling and separation processing to integrate image and text data. Additionally, we suggest the Dynamic Loss Weight Allocation (DLWA) method for multi-task learning to refine the training procedure. We confirm the superiority of the proposed method on the publicly available multi-modal skin lesion dataset, Derm7pt. The average accuracy of multi-modal skin disease classification is 79.14%, surpassing the current state-of-the-art methods.

关键词： Training Attention mechanisms Accuracy Fuses Multitasking Dynamic scheduling Skin Resource management Lesions Diseases

来源：评论

学校读者我要写书评

暂无评论

Timing-Driven High-Level Synthesis for Continuous-Flow Microfluidic Biochips

Timing-Driven High-Level Synthesis for Continuous-Flow Micro...

引用

IEEE International Symposium on Quality Electronic Design

作者： Zhengyang Ye Zhisheng Chen Youlin Pan Genggeng Liu Wenzhong Guo Tsung-Yi Ho Xing Huang College of Computer and Data Science Fuzhou University Fuzhou China Engineering Research Center of Big Data Intelligence Ministry of Education Fuzhou China Fujian Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University Fuzhou China School of Informatics Xiamen University Xiamen China Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong China School of Computer Science Northwestern Polytechnical University Xi’an China

ISBN: (数字)9798350309270

ISBN: (纸本)9798350309287

As continuous-flow microfluidic biochips technology rapidly advances, its automated synthesis methods require adapting to increasingly complex and stringent biochemical applications. In particular, with the introduction of time-sensitive bioassays, in this work, we present a timing-driven high-level synthesis method to optimize the resource efficiency of bioassays considering the time constraints. Specifically, we investigate the existing forms of time constraints and propose a corresponding flexible modeling approach based on reaction-control which can achieve downward compatibility for physical design. Additionally, we also analyze the impact of storage and wash on time constraints. The experimental results show that our method achieves efficient execution and low devices cost while satisfying time constraints.

关键词： Costs Biochips Time factors Microfluidics Physical design

来源：评论

学校读者我要写书评

暂无评论

Unleashing the Power of Federated Learning in Fragmented Digital Healthcare Systems: A Visionary Perspective

Unleashing the Power of Federated Learning in Fragmented Dig...

引用

International Conference on Software, Knowledge Information, Industrial Management and Applications (SKIMA)

作者： Marzia Hoque Tania David A. Clifton Centre for Big Data Research in Health University of New South Wales Sydney Australia Department of Engineering Science Institute of Biomedical Engineering University of Oxford Oxford United Kingdom Oxford Centre for Advanced Research (OSCAR) University of Oxford Suzhou China

Digital healthcare landscape, including infrastructure, governance, interoperability, and user adoption, are continuously evolving, some taking more centralised approach, while others with higher degree of fragmentation. Attitude towards centralised healthcare systems in affluent countries are primarily influenced by historical development, infrastructure investments, and regulatory frameworks, which offers advantages with respect to standardised practises, centralised decision making, and economies of scale. In contrast, complexities due to diverse stakeholders, interoperability challenges, privacy and security concerns often pose challenges in achieving a completely centralised healthcare system even in high income countries such as the United Kingdom or in federal systems such as the United States. Moreover, decentralised healthcare systems are more prevalent in resource-poor countries. This paper presents our viewpoint and perspectives on the potential of federated learning in decentralised healthcare systems, especially in countries with infrastructure constraints and discusses its advantages, privacy and security concerns, and challenges. As data-hungry artificial intelligence-enabled systems are gradually changing the healthcare ecosystem, federated learning presents an opportunity for distributing the machine learning training process across multiple decentralised edge devices with reduced data transfer. Therefore, the decentralised digital healthcare system can leverage the collaborative model training while protecting highly sensitive and personal health information. However, challenges related to data heterogeneity, communication latency, and model aggregation need to be addressed for successful implementation of such systems. Adapting the federated learning framework to the specific needs and constraints of low and middle-income countries is crucial to unlock its potential in improving healthcare outcomes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Remote Sensing Object Classification and Extraction Method Based on Improved U-Net 1

A Remote Sensing Object Classification and Extraction Method...

引用

1st IEEE International Conference on data Intelligence and Innovative Application, DIIA 2024

作者： Wen, Lili Qu, Hui Wu, Man Lai, Junxiang Zhang, Wei Research Center for Carbon Sink and Low-Carbon Engineering Beibu Gulf of Guangxi Guangxi Academy of Marine Sciences line Nanning China National Marine Data and Information Service Tianjin China School of Electrical Engineering Guangxi University Nanning China Guangxi Academy of Sciences Guangxi Key Laboratory of Marine Environmental Science Nanning China Electric Power Research Institute of Guangxi Power Grid Co. Ltd. Guangxi Key Laboratory of Intelligent Control and Operation and Maintenance of Power Equipment Nanning China

ISBN: (纸本)9798331504434

This paper addresses the problems of low accuracy and poor adaptability to complex scenes in traditional remote sensing object classification methods, and proposes a remote sensing object classification extraction method based on an improved U-Net. We retained the U-shaped symmetrical structure of U-Net and introduced the Swin Transformer block with self-attention mechanism to replace the original convolutional operation, enhancing the model's ability to extract global features. We also constructed a Bottleneck structure between the encoder and decoder to learn deep feature representations and form a new U-shaped network based on Transformer, improving the network's image semantic segmentation ability in scenarios with small targets and complexity. Experimental results show that the method proposed in this paper has an improvement of more than 9% in evaluation indicators such as recognition accuracy and recall rate compared with existing U-Net and Res-UNet methods, has high semantic segmentation accuracy and robustness, and provides new ideas and technical support for remote sensing image object classification extraction. © 2024 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Regularized Multivariate Functional Principal Component Analysis

arXiv

引用

arXiv 2023年

作者： Haghbin, Hossein Zhao, Yue Maadooliat, Mehdi Artificial Intelligence and Data Mining Research Group ICT Research Institute Faculty of Intelligent Systems Engineering and Data Science Persian Gulf University Bushehr Iran Department of Mathematical and Statistical Sciences Marquette University MilwaukeeWI53233 United States

Multivariate Functional Principal Component Analysis (MFPCA) is a valuable tool for exploring relationships and identifying shared patterns of variation in multivariate functional data. However, controlling the roughness of the extracted Principal Components (PCs) can be challenging. This paper introduces a novel approach called regularized MFPCA (ReMFPCA) to address this issue and enhance the smoothness and interpretability of the multivariate functional PCs. ReMFPCA incorporates a roughness penalty within a penalized framework, using a parameter vector to regulate the smoothness of each functional variable. The proposed method generates smoothed multivariate functional PCs, providing a concise and interpretable representation of the data. Extensive simulations and real data examples demonstrate the effectiveness of ReMFPCA and its superiority over alternative methods. The proposed approach opens new avenues for analyzing and uncovering relationships in complex multivariate functional *** Codes 62Hxx, 46-xx (Primary) 62H25 (Secondary) © 2023, CC BY.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Character Spray Code Detection of Steel Coil Based on High Generalization ABCNet

Character Spray Code Detection of Steel Coil Based on High G...

引用

Electronic Information engineering and Computer science (EIECS), 2021 International Conference on

作者： Hongyang Zhou Xinyue Mao Ziyang Wu Yuan Zhang College of Big Data and Artificial Intelligence China Telecom Research Institute Shanghai China University of California Irvine Irvine USA College of Information Science and Electronic Engineering Zhejiang University Hangzhou China

In order to improve the accuracy of end-to-end text detection algorithm, a steel coil inkjet detection algorithm based on high generalization ABCNet is proposed. In the mainstream ABCNet text detection algorithm, the Backbone module is optimized and the high generalization ability is integrated to improve the detection accuracy. Experiments show that the algorithm can effectively improve the detection accuracy of steel coil inkjet coding, and compared with EAST, TextSnake, DBNet + +, PAN-640 and other models in terms of accuracy Precision and F-Measure. The results show that the algorithm has obvious advantages. On the steel coil detection dataset, Precision and F-Measure achieved good results, Precision was 99.7%, and F-Measure was 99.2%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fil-Cad: A Fault Detection Method for Iiot Based on Federated Incremental Learning with Class Accuracy Distillation

SSRN

引用

SSRN 2024年

作者： Liu, Yanhua Fang, Wenyu Huang, Wei Wang, Xiaofeng Zhao, Baokang Liu, Ximeng College of Computer and Data Science Fuzhou University Fuzhou350108 China Engineering Research Center of Big Data Intelligence Ministry of Education Fuzhou350108 China Fujian Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University Fuzhou350108 China College of Computer National University of Defense Technology Changsha410073 China

Fault detection is a research topic attracting much attention in IIoT. However, the continuous accumulation and growth of IIoT data leads to the catastrophic forgetting problem in the fault detection model and data heterogeneity, which creates a bottleneck for the performance and generalization of federated learning in fault detection. To improve the fault detection performance of federated learning models in data incremental scenarios, we propose a federatedincremental-learning-based fault detection method for IIoT, namely FIL-CAD. Firstly, comparing the knowledge distillation, the accuracies of different data classes in the global and local models are chosen as the model lightweight learning features, and the loss function is weighted. Secondly, we introduce the concept of data learning depth. It computes different model weights in the aggregation based on their degrees of data sample learning and model training participation and obtains the global model through weighted aggregation. Additionally, FIL-CAD supports the combination of differential privacy to protect data privacy. The proposed method is compared with four benchmark methods on four datasets, which include CWRU and MFPT bearing datasets. Then, the experimental results show that FIL-CAD performs optimally in accuracy and F1-Score, and the average accuracy is improved by 1.32%-18.52%, demonstrating the superiority of FIL-CAD. © 2024, The Authors. All rights reserved.

关键词： Fault detection

来源：评论

学校读者我要写书评

暂无评论

Prototype and Instance Contrastive Learning for Unsupervised Domain Adaptation in Speaker Verification

arXiv

引用

arXiv 2024年

作者： Huang, Wen Han, Bing Chen, Zhengyang Wang, Shuai Qian, Yanmin Auditory Cognition and Computational Acoustics Lab MoE Key Lab of Artificial Intelligence AI Institute Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Shenzhen China

Speaker verification system trained on one domain usually suffers performance degradation when applied to another domain. To address this challenge, researchers commonly use feature distribution matching-based methods in unsupervised domain adaptation scenarios where some unlabeled target domain data is available. However, these methods often have limited performance improvement and lack generalization in various mismatch situations. In this paper, we propose Prototype and Instance Contrastive Learning (PICL), a novel method for unsupervised domain adaptation in speaker verification through dual-level contrastive learning. For prototype contrastive learning, we generate pseudo labels via clustering to create dynamically updated prototype representations, aligning instances with their corresponding class or cluster prototypes. For instance contrastive learning, we minimize the distance between different views or augmentations of the same instance, ensuring robust and invariant representations resilient to variations like noise. This dual-level approach provides both high-level and low-level supervision, leading to improved generalization and robustness of the speaker verification model. Unlike previous studies that only evaluated mismatches in one situation, we have conducted relevant explorations on various datasets and achieved state-of-the-art performance currently, which also proves the generalization of our method. Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

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：