检索结果-内蒙古大学图书馆

32nd International Joint Conference on Artificial Intelligence, IJCAI 2023

作者： Liu, Shiyu Wei, Linsen Lv, Shaogao Li, Ming University of Electronic Science and Technology of China China School of Astronautics Northwestern Polytechnical University China Department of Statistics and Data Science Nanjing Audit University China Key Laboratory of Intelligent Education Technology Application of Zhejiang Province Zhejiang Normal University China

ISBN: (纸本)9781956792034

Graph convolutional networks (GCN) are viewed as one of the most popular representations among the variants of graph neural networks over graph data and have shown powerful performance in empirical experiments. That 2-based graph smoothing enforces the global smoothness of GCN, while (soft) 1-based sparse graph learning tends to promote signal sparsity to trade for discontinuity. This paper aims to quantify the trade-off of GCN between smoothness and sparsity, with the help of a general p-regularized (1 p-regularized learning scheme does not satisfy such a smooth condition. To tackle this issue, we propose a novel SGD proximal algorithm for GCNs with an inexact operator. For a single-layer GCN, we establish an explicit theoretical understanding of GCN with the p-regularized stochastic learning by analyzing the stability of our SGD proximal algorithm. We conduct multiple empirical experiments to validate our theoretical findings. © 2023 International Joint Conferences on Artificial Intelligence. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Mixture of Experts as Representation Learner for Deep Multi-View Clustering 39

Mixture of Experts as Representation Learner for Deep Multi-...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Zhang, Yunhe Cai, Jinyu Wu, Zhihao Wang, Pengyang Ng, See-Kiong Department of Computer and Information Science SKL-IOTSC University of Macau China Institute of Data Science National University of Singapore Singapore College of Computer Science and Technology Zhejiang University China

ISBN: (纸本)157735897X

Multi-view clustering (MVC) aims to integrate information from diverse data sources to facilitate the clustering process, which has achieved considerable success in various real-world applications. However, previous MVC methods typically employ one of two strategies: (1) designing separate feature extraction pipelines for each view, which restricts their ability to fully exploit collaborative potential;or (2) employing a single shared representation module, which hinders the capture of diverse, view-specific representations. To tackle these challenges, we introduce Deep Multi-View Clustering via Collaborative Experts (DMVC-CE), a novel MVC approach that employs the Mixture of Experts (MoE) framework. DMVC-CE incorporates a gating network that dynamically selects multiple experts for handling each data sample, capturing diverse and complementary information from different views. Additionally, to ensure balanced expert utilization and maintain their diversity, we introduce an equilibrium loss and a multi-expert distinctiveness enhancer. The equilibrium loss prevents excessive reliance on specific experts, while the distinctiveness enhancer encourages each expert to specialize in different aspects of the data, thereby promoting diversity in learned representations. Comprehensive experiments on various multi-view benchmark datasets demonstrate the superiority of DMVC-CE compared to state-of-the-art MVC baselines. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

HIGHER CONNECTIVITY OF THE MORSE COMPLEX

引用

Proceedings of the American Mathematical Society, Series B 2022年第14期9卷 135-149页

作者： Scoville, Nicholas A. Zaremsky, Matthew C. B. Department of Mathematics and Computer Science Ursinus College Collegeville 19426 PA United States Department of Mathematics and Statistics University at Albany (SUNY) Albany 12222 NY United States

The Morse complex M(Δ) of a finite simplicial complex Δ is the complex of all gradient vector fields on Δ. In this paper we study higher connectivity properties of M(Δ). For example, we prove that M(Δ) gets arbitrarily highly connected as the maximum degree of a vertex of Δ goes to ∞, and for Δ a graph additionally as the number of edges goes to ∞. We also classify precisely when M(Δ) is connected or simply connected. Our main tool is Bestvina–Brady Morse theory, applied to a “generalized Morse complex.”. © 2022, American Mathematical Society. All rights reserved.

关键词： discrete Morse theory higher connectivity Morse complex

来源：评论

学校读者我要写书评

暂无评论

Learning Relational Tabular data without Shared Features

arXiv

引用

arXiv 2025年

作者： Wu, Zhaomin Wang, Shida Wang, Ziyang He, Bingsheng Institute of Data Science National University of Singapore Singapore Department of Mathematics National University of Singapore Singapore Department of Computer Science National University of Singapore Singapore

Learning relational tabular data has gained significant attention recently, but most studies focus on single tables, overlooking the potential of cross-table learning. Cross-table learning, especially in scenarios where tables lack shared features and pre-aligned data, offers vast opportunities but also introduces substantial challenges. The alignment space is immense, and determining accurate alignments between tables is highly complex. We propose Latent Entity Alignment Learning (Leal), a novel framework enabling effective cross-table training without requiring shared features or pre-aligned data. Leal operates on the principle that properly aligned data yield lower loss than misaligned data, a concept embodied in its soft alignment mechanism. This mechanism is coupled with a differentiable cluster sampler module, ensuring efficient scaling to large relational tables. Furthermore, we provide a theoretical proof of the cluster sampler’s approximation capacity. Extensive experiments on five real-world and five synthetic datasets show that Leal achieves up to a 26.8% improvement in predictive performance compared to state-of-the-art methods, demonstrating its effectiveness and scalability. © 2025, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

The broken sample problem revisited: Proof of a conjecture by Bai-Hsing and high-dimensional extensions

arXiv

引用

arXiv 2025年

作者： Jiao, Simiao Wu, Yihong Xu, Jiaming Fuqua School of Business Duke University United States Department of Statistics and Data Science Yale University United States

We revisit the classical broken sample problem: Two samples of i.i.d. data points X = {X1, . . ., Xn} and Y = {Y1, . . ., Ym} are observed without correspondence with m ≤ n. Under the null hypothesis, X and Y are independent. Under the alternative hypothesis, Y is correlated with a random subsample of X, in the sense that (Xπ(i), Yi)’s are drawn independently from some bivariate distribution for some latent injection π: [m] → [n]. Originally introduced by DeGroot, Feder, and Goel [7] to model matching records in census data, this problem has recently gained renewed interest due to its applications in data de-anonymization, data integration, and target tracking. Despite extensive research over the past decades, determining the precise detection threshold has remained an open problem even for equal sample sizes (m = n). Assuming m and n grow proportionally, we show that the sharp threshold is given by a spectral and an L2 condition of the likelihood ratio operator, resolving a conjecture of Bai and Hsing [2] in the positive. These results are extended to high dimensions and settle the sharp detection thresholds for Gaussian and Bernoulli models. Copyright © 2025, The Authors. All rights reserved.

关键词： data integration

来源：评论

学校读者我要写书评

暂无评论

Towards Real Time Traffic Analysis with Reduced Image Quality

Towards Real Time Traffic Analysis with Reduced Image Qualit...

引用

IEEE India Conference (INDICON)

作者： Rohit Mohanraj Erunika Oshani Dayaratna Department of Electronics and Communication Engineering National Institute of Technology Warangal Warangal India Department of Statistics and Computer Science Faculty or science University of peradeniya SriLanka

ISBN: (数字)9798350391282

ISBN: (纸本)9798350391299

This paper presents a detailed study of two computationally efficient object recognition models-YOLOv5 and Fast R-CNN with a MobileNet backbone-focusing on their performance under different degrees of image / video qualities. Blurring is a common problem in autonomous driving due to motion, weather conditions, and other factors, which can greatly reduce picture quality and, consequently, the accuracy of object recognition. Also, the low-quality images produced by on-car devices suffer when trying to utilize them for advanced detection applications. Utilizing the Zenseact Open dataset (ZOD), images are generated by varying kernel sizes of Gaussian blur to examine how each model's recognition accuracy is affected by picture degradation. Our study shows key differences in the stability of these models, showing the trade-offs between processing speed and recognition confidence in poor visual settings. The results provide helpful guidance for choosing object recognition systems that keep performance in difficult situations, crucial for the safety and efficiency of driverless cars.

关键词： YOLO Performance evaluation Degradation Training Adaptation models Visualization Accuracy Real-time systems Object recognition Video recording

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Approach for Hepatic Lesion Detection

Deep Learning Approach for Hepatic Lesion Detection

引用

Intelligent Methods, Systems, and Applications (IMSA)

作者： Manal Makram Ammar Mohammed Department of Data Science FGSSR Cairo University Cairo Egypt Computer Science FGSSR Cairo University Cairo Egypt

ISBN: (数字)9798350362633

ISBN: (纸本)9798350362640

Egypt ranks second worldwide for liver cancer mortality, with 4.57% of global hepatic tumor deaths. Liver disease-related fatalities make up 11.20% of Egypt's total mortality. Additionally, Egypt is among the top 10 countries for obesity, a leading cause of metabolic-associated fatty liver disease. Liver cancer malignancy varies, with the diagnosis and treatment of these differences crucial for patient survival rates and lifespans. Thus, early detection of liver tumors is essential to successful treatment and better patient outcomes. This paper introduces a multiview liver dataset sourced from Ain Shams University Specialized Hospital (ASUSH) in Egypt and annotated by expert radiologists. The dataset includes 280 patients with 14,096 computed tomography (CT) images. This study's ASUSH dataset is groundbreaking, representing Egypt's first multi-class liver cancer dataset with eight distinct classes. Also, furthermore, the paper benchmarks the dataset to multiple deep learning models, including a proposed CNN, as well as fine-tuned pretrained models like VGG16, VGG19, ResNet50, ResNet101, EfficientNetB1, EfficientNetB2, Xception, Vision Transformer (ViT), Inception-V3, and InceptionRestNetV2. Notably, employing an ensemble learning approach significantly enhanced the results for view two. The experimental results on several performance metrics reveal that the models achieved an accuracy of 98% in predicting liver cancer variants.

关键词： Deep learning Measurement Liver cancer Obesity Liver diseases Computed tomography Computational modeling

来源：评论

学校读者我要写书评

暂无评论

Self-Sovereign Identity in Semi-Permissioned Blockchain Networks Leveraging Ethereum and Hyperledger Fabric

Self-Sovereign Identity in Semi-Permissioned Blockchain Netw...

引用

Digital Health (ICDH), IEEE International Conference on

作者： Niranjan Sahi Anda Liang Wyn Van Devanter Konstantinos Oikonomou Peng Zhang Department of Computer Science Vanderbilt University Nashville TN USA Department of Computer Science and Data Science Institute Vanderbilt University Nashville TN USA

Patients often have their healthcare data stored in centralized systems, leading to challenges when reconciling or consolidating their data across providers due to centralized databases that store patient identities. The challenges disrupt the flow of patient care where time is sensitive for both patients and providers. Decentralized technologies have enabled a new identity model–Self-Sovereign Identity (SSI)–that grants individuals the right to freely control, access, and share their own data. This work proposes a system that achieves SSI in a semi-permissioned blockchain network using an open protocol as the certificate of authority and several guidelines for securely handling transactions in the network. Open protocols like Keccak can grant access to a permission-based network such as Hyperledger Fabric. The network architecture ensures data security and privacy through mechanisms of multi-signature transactions and guidelines for storing transactions locally, making this architecture ideal for privacy-centered use cases, such as healthcare data-sharing applications. The ultimate goal is to give patients full control over their identity and other data derived from their identity within a semi-permissioned network.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Poster: Extracting and Annotating Mental Health Forum Corpus: A Comprehensive Validation Pipeline

Poster: Extracting and Annotating Mental Health Forum Corpus...

引用

IEEE International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE)

作者： Rohith Sundar Jonnalagadda Abm Adnan Azmee Dinesh Attota Md Abdullah Al Hafiz Khan Yong Pei Monica Nandan Department of Computer Science Data Science and Analytics Social Work and Human Services Kennesaw State University Marietta USA

ISBN: (数字)9798350345018

ISBN: (纸本)9798350345025

This research emphasizes the essential role of mental health forums as vital online communities providing solace, support, and resources for individuals grappling with mental health issues, especially among young people. Acknowledging the presence of severe content in some posts, indicative of acute distress and potential self-harm risk, the study draws on prior research highlighting the forums' critical role in fulfilling lower-level support needs for young individuals. By employing advanced classification and summarization techniques, namely Long short-term memory (LSTM), Bidirectional LSTM (BiLSTM) and BERT, the project aims to enhance the efficiency of these forums through systematic categorization and summarization of user posts. Preliminary results show promising outcomes, with improved post-classification accuracy ranging from 40% to 83.33% and average Rouge F1 scores ranging from 43% to 54%. This research contributes to fortifying the role of mental health forums in providing essential support to young individuals in distress.

关键词： Systematics Accuracy Pipelines Mental health Bidirectional control Distance measurement Encoding

来源：评论

学校读者我要写书评

暂无评论

Advancing Healthcare IoT: Blockchain and Federated Learning Integration for Enhanced Security and Insights

Advancing Healthcare IoT: Blockchain and Federated Learning ...

引用

2024 International Conference on Communication, computer sciences and Engineering, IC3SE 2024

作者： Malik, Rida Ur-Rehaman, Atta Razzaq, Hamza Bhatt, Chandradeep Kaushik, Keshav Khan, Inam Ullah Riphah Institute Of System Engineering Department Of Information Security Islamabad Pakistan Riphah Institute Of System Engineering Department Of Data Science Islamabad Pakistan Graphic Era Hill University Department Of Cse Uttarakhand India Graphic Era Deemed To Be University Uttarakhand Dehradun India School Of Computer Science University Of Petroleum And Energy Studies Dehradun India Seas Isra University Department Of Electronic Engineering Islamabad Campus Pakistan

ISBN: (纸本)9798350366846

Fast-growing IoT devices in healthcare have ushered in a new era of data-driven patient care and treatment. These gadgets capture too much sensitive medical data, a major issue. We need to boost data analysis, security, and privacy. This research uses cutting-edge concepts like federated learning and blockchain technology to improve Healthcare IoT analytics and security. This study examines blockchain and federated learning's numerous integration aspects. It begins with data security, employing blockchain's immutability and decentralization to prevent manipulation. Federation learning was utilized to solve privacy concerns while adhering to tight healthcare data security regulations, providing strong data analysis without releasing raw, patient-specific information. This study examines how federated learning and blockchain technologies enhance healthcare IoT systems' analytics. This connection allows safe, collaborative analysis among healthcare professionals and devices, offering more insights for research, diagnostics, and tailored patient care. How to integrate these cutting-edge technologies into healthcare IoT networks is also explored. © 2024 IEEE.

关键词： Diagnosis

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：