检索结果-内蒙古大学图书馆

IEEE Congress on Cybermatics: 17th IEEE International Conference on Internet of Things, iThings 2024, 20th IEEE International Conference on Green Computing and Communications, GreenCom 2024, 17th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2024, 10th IEEE International Conference on Smart Data, SmartData 2024

作者： Chen, Shiyu Li, Qianmu Nanjing University of Science and Technology School of Computer Science and Engineering Nanjing China

ISBN: (纸本)9798350351637

Social influence is an individual's ability to change the thoughts or behaviors of others due to factors such as social status, social connections, and social wealth. Studying social influence, especially modeling and measuring the influence between people on large-scale social network data, is very important for product promotion, friend recommendation, expert discovery and user behavior prediction in social applications. In addition, it plays an important role in promoting the virtual business market and national public opinion monitoring. From a computational point of view, the research on social influence includes detection, modeling, measurement, and application of social influence. First, a training dataset is generated through independent cascade and linear threshold models. The influence of each node in the network is then calculated through graph convolutional network and reinforcement learning techniques. Finally, the proposed model is tested on three real-world large-scale datasets. Experimental results show that the proposed method can help predict the influence of nodes in social networks. © 2024 IEEE.

关键词： deep learning graph convolutional network graph embedding reinforcement learning social influence social networks

来源：评论

学校读者我要写书评

暂无评论

Collaborative Learning with Curriculum Loss for Accurate and Interpretable Weakly Supervised Whole-Slide Image Classification Collaborative Learning with Curriculum Loss for Whole-Slide Image Classification

Collaborative Learning with Curriculum Loss for Accurate and...

引用

2024 International Conference on Machine Learning and Intelligent Computing, MLIC 2024

作者： Yang, Rui Liu, Pei Ji, Luping School of Computer Science and Engineering University of Electronic Science and Technology of China China

Weakly-Supervised Learning (WSL) has been increasingly concerned in Whole-Slide Image (WSI) classification, meanwhile, an open question arises: could WSL-based models provide us with an accurate interpretation of their decisions? Although many research works have made exciting progress via building an Auxiliary Instance Branch (AIB) on a bag-level network, there are still two typical problems to be confronted with in training WSL-based AIB: i) an overwhelming influence of negative instances and ii) the inconsistent learning between bag-level network and AIB. To address them, this paper proposes collaborative learning with curriculum loss. This scheme, on one hand, provides a curriculum loss for optimizing AIB, to alleviate the first problem. Considering the knowledge reliability in WSL, this loss generalizes an original quality focal loss to WSL scenarios by curriculum instances. On the other hand, to overcome the second problem, this scheme trains a bag-level network under the supervision of AIB by a reversed curriculum loss, making both learn collaboratively. Comparative experiments prove that our scheme could often surpass existing ones in both accuracy and interpretability. Moreover, it is found that the knowledge reliability-inspired curriculum instance is a critical factor in bringing comprehensive improvements. © 2024 R. Yang, P. Liu & L. Ji.

关键词： Computational Pathology Curriculum Loss Model Interpretability Multiple Instance Learning Weakly-Supervised Learning

来源：评论

学校读者我要写书评

暂无评论

STLL: Spatio-Temporal Graph and Lane Sequence Based Lightweight Vehicle Trajectory Prediction 10

STLL: Spatio-Temporal Graph and Lane Sequence Based Lightwei...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Liu, Yuzhen Huang, Lei Wang, Xiaoliang Ai, Yuanyuan Zhou, Xiaolan Hunan University of Science and Technology School of Computer Science and Engineering Xiangtan China

ISBN: (纸本)9798331520861

Accurate and swift prediction of surrounding vehicle trajectories is essential for autonomous driving. Currently, numerous methods have achieved excellent accuracy in trajectory prediction but they often overlook the prediction speed. Autonomous vehicles must quickly respond to various situations for safe driving. Therefore, we propose a lightweight model named STLL. We model the interactions among vehicles using a graph that fully considers the spatio-temporal information. Additionally, we represent all lanes by a sequence and employ an efficient method to encode lanes. These approaches effectively enable the model to extract and fuse features of vehicles and lanes, leading to accurate and swift trajectory predictions. We conduct our experiments on the widely recognized real-world trajectory prediction datasets: nuScenes and Argoverse. The results demonstrate that our method not only significantly reduces prediction time and the number of model parameters but also maintains high prediction accuracy. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Graph Convolutional Networks with Fine-Tuned Word Representations for Visual Question Answering 10

Graph Convolutional Networks with Fine-Tuned Word Representa...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Owusu, Jim Wilson Zakari, Rufai Yusuf Qin, Ke He, Tao University of Electronic Science and Technology of China School of Computer Science and Engineering China

ISBN: (纸本)9798331520861

Recent advances in graph convolutional networks (GCNs) have demonstrated their effectiveness in vision-language tasks such as visual question answering (VQA), primarily due to their ability to capture both spatial and semantic relationships. However, through a series of experiments applying GCNs to VQA tasks, we observe that the performance of GCNs is highly sensitive to various subtasks, often resulting in unstable outputs. This variability can be attributed to factors such as the size of the training and testing datasets, evaluation metrics, and the selection of hyperparameters. To ensure fair evaluations of GCN-based VQA models, it is essential to implement consistent evaluation schemes across these factors. In this study, we propose a GCN framework for VQA that leverages fine-tuned word representations from the GloVe model, specifically tailored to address reasoning-based questions. We comprehensively assess the performance of our framework using multiple evaluation metrics on both the GQA and VQA 2.0 datasets. Our experimental results demonstrate significant improvements over existing methods, showcasing the effectiveness of the proposed approach in evaluation setting. © 2024 IEEE.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

An attention-based teacher-student model for multivariate short-term landslide displacement prediction incorporating weather forecast data

引用

Journal of Mountain science 2024年第8期21卷 2739-2753页

作者： CHEN Jun HU Wang ZHANG Yu QIU Hongzhi WANG Renchao Sichuan Artificial Intelligence Research Institute Yibin 644000China School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu 611731China School of Architecture and Civil Engineering Chengdu UniversityChengdu 610106China

Predicting the displacement of landslide is of utmost practical importance as the landslide can pose serious threats to both human life and ***,traditional methods have the limitation of random selection in sliding window selection and seldom incorporate weather forecast data for displacement prediction,while a single structural model cannot handle input sequences of different lengths at the same *** order to solve these limitations,in this study,a new approach is proposed that utilizes weather forecast data and incorporates the maximum information coefficient(MIC),long short-term memory network(LSTM),and attention mechanism to establish a teacher-student coupling model with parallel structure for short-term landslide displacement *** MIC,a suitable input sequence length is selected for the LSTM *** investigate the influence of rainfall on landslides during different seasons,a parallel teacher-student coupling model is developed that is able to learn sequential information from various time series of different *** teacher model learns sequence information from rainfall intensity time series while incorporating reliable short-term weather forecast data from platforms such as China Meteorological Administration(CMA)and Reliable Prognosis(https://***)to improve the model’s expression capability,and the student model learns sequence information from other time *** attention module is then designed to integrate different sequence information to derive a context vector,representing seasonal temporal attention ***,the predicted displacement is obtained through a linear *** proposed method demonstrates superior prediction accuracies,surpassing those of the support vector machine(SVM),LSTM,recurrent neural network(RNN),temporal convolutional network(TCN),and LSTM-Attention *** achieves a mean absolute error(MAE)of 0.072 mm,root mean square error(RMSE)of 0.096 mm,and pearson correlation coefficients(PC

关键词： Landslide prediction MIC LSTM Attention mechanism Teacher Student model Prediction stability and interpretability

来源：评论

学校读者我要写书评

暂无评论

A Unimodal Valence-Arousal Driven Contrastive Learning Framework for Multimodal Multi-Label Emotion Recognition 24

A Unimodal Valence-Arousal Driven Contrastive Learning Frame...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Zheng, Wenjie Yu, Jianfei Xia, Rui School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China

ISBN: (纸本)9798400706868

Multimodal Multi-Label Emotion Recognition (MMER) aims to identify one or more emotion categories expressed by an utterance of a speaker. Despite obtaining promising results, previous studies on MMER represent each emotion category using a one-hot vector and ignore the intrinsic relations between emotions. Moreover, existing works mainly learn the unimodal representation based on the multimodal supervision signal of a single sample, failing to explicitly capture the unique emotional state of each modality as well as its emotional correlation between samples. To overcome these issues, we propose a Unimodal Valence-Arousal driven contrastive learning framework (UniVA) for the MMER task. Specifically, we adopt the valence-arousal (VA) space to represent each emotion category and regard the emotion correlation in the VA space as priors to learn the emotion category representation. Moreover, we employ pre-trained unimodal VA models to obtain the VA scores for each modality of the training samples, and then leverage the VA scores to construct positive and negative samples, followed by applying supervised contrastive learning to learn the VA-aware unimodal representations for multi-label emotion prediction. Experimental results on two benchmark datasets MOSEI and M3ED show that the proposed UniVA framework consistently outperforms a number of existing methods for the MMER task. The source code is publicly released at https://***/NUSTM/UniVA. © 2024 ACM.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Simplified Kernel-Based Cost-Sensitive Broad Learning System for Imbalanced Fault Diagnosis

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第12期5卷 6629-6644页

作者： Yang, Kaixiang Chen, Wuxing Shi, Yifan Yu, Zhiwen Chen, C. L. Philip South China University of Technology School of Computer Science and Engineering Guangdong Guangzhou510006 China South China University of Technology School of Future Technology Guangdong Guangzhou510006 China Pengcheng Laboratory Guangdong Shenzhen518066 China Huaqiao University College of Engineering Fujian Quanzhou362021 China South China University of Technology School of Computer Science and Engineering Guangdong Guangzhou510650 China

In the field of intelligent manufacturing, tackling the classification challenges caused by imbalanced data is crucial. Although the broad learning system (BLS) has been recognized as an effective and efficient method, its performance wanes with imbalanced datasets. Therefore, this article proposes a novel approach named simplified kernel-based cost-sensitive broad learning system (SKCSBLS) to address these issues. Based on the framework of cost-sensitive broad learning system (CSBLS) that assigns distinctive adjustment costs for individual classes, SKCSBLS emphasizes the importance of the minority class while mitigating the impact of data imbalance. Additionally, considering the complexity introduced by noisy or overlapping data points, we incorporate kernel mapping into the CSBLS. This improvement not only improves the system's capability to handle overlapping classes of samples, but also improves the overall classification effectiveness. Our experimental results highlight the potential of SKCSBLS in overcoming the challenges inherent in unbalanced data, providing a robust solution for advanced fault diagnosis in intelligent systems. © 2020 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

A lightweight network intrusion detection model based on convolution and attention mechanisms 3

A lightweight network intrusion detection model based on con...

引用

3rd International Conference on Electronic Information engineering, Big Data, and computer technology, EIBDCT 2024

作者： Fang, Zizheng Liu, Yaodi Yuan, Shuai Ye, Tiancheng School of Cyber Science and Engineering Nanjing University of Science and Technology Nanjing China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China

ISBN: (纸本)9781510680449

With the development of the Internet of Things (IoT), people's lives have become more intelligent. However, the proliferation of numerous small devices has also introduced serious risks to network security. Network intrusion detection, as a proactive defense technique, can effectively address network attacks. However, IoT devices are constrained by limited storage and computational resources, and many current intrusion detection models suffer from an excessive number of parameters. Therefore, this thesis proposes a network intrusion detection model based on depthwise separable convolution (DSC) and a convolutional block attention mechanism (CBAM). A convolution with an inverse residual structure (IRCNN) is designed for network intrusion detection based on DSC. Additionally, a convolutional block attention mechanism is introduced in the feature extraction phase to enhance the model's ability to capture important features. To validate the performance of the proposed model in this paper, experimental validation is conducted on the CIC-IDS-2017 dataset. The results demonstrate that the IRCNN-CBAM model exhibits high detection performance. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Supervised Data-Dependent Kernel Low-Rank Preserving Projection for Image Recognition 5

Supervised Data-Dependent Kernel Low-Rank Preserving Project...

引用

5th International Conference on Intelligent Computing and Human-computer Interaction, ICHCI 2024

作者： Li, Linhao Zhang, Ao Jiangsu University of Science and Technology School of Computer Science and Engineering Zhenjiang China

ISBN: (纸本)9798350368284

Dimensionality reduction (DR) is a frequently used method to handle data with many dimensions. This article presents a novel approach to nonlinear dimensionality reduction, referred to as supervised data-dependent kernel low-rank preserving projection (SDKLRPP). Designed for image recognition tasks. The method enhances performance by utilizing a data-dependent kernel (DK). Unlike traditional nonlinear dimensionality reduction techniques, the SDKLRPP method described in this paper can fine-Tune the coefficients within the DK and uncover The paper elucidates the low-rank reconstructive relationship inherent in the data. It also illustrates how the manifold structure within the feature space is mirrored in the label space, grounded in the principle of smoothness. This enables label information to direct the optimization of coefficients within the DK, thereby enhancing the capacity to discern among various classes. Upon attaining the optimal low-rank reconstructive relationship, a transformation matrix is computed to maintain this relationship, which is subsequently utilized to project the data into a reduced-dimensional space. The efficacy of the proposed SDKLRPP methodology is substantiated through experiments conducted on five extensively acknowledged databases. © 2024 IEEE.

关键词： Dimensionality reduction

来源：评论

学校读者我要写书评

暂无评论

Self-Supervised Learning Recommendation with Enhanced Long-Tail Nodes 5

Self-Supervised Learning Recommendation with Enhanced Long-T...

引用

5th International Conference on Artificial Intelligence and computer engineering, ICAICE 2024

作者： Zhao, Zhiyu Wang, Yongli Liu, Dongmei School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China

ISBN: (纸本)9798331528911

Achieving a balance between recommending popular items and long-tail items has consistently posed a challenge in the field of recommendation systems. Traditional recommendation algorithms often exhibit a bias towards recommending popular items, resulting in limited exposure for long-tail items. To address this issue, our study introduces an innovative recommendation algorithm based on graph contrastive learning. This algorithm incorporates a dual-view learning mechanism with the objective of increasing the exposure of long-tail items while ensuring the accuracy of recommendations. Two distinct views are generated: one captures in-depth information of long-tail nodes through random walk strategies, while the other utilizes Graph Attention Networks (GAT) to acquire node importance and contextual information. By means of contrastive learning with these two views, our model attains comprehensive node representations. The study strives to achieve a balanced learning process between long-tail and high-frequency nodes. Experimental results conducted on multiple recommendation system datasets substantiate that our approach significantly enhances the exposure of long-tail items compared to existing recommendation algorithms, while maintaining high accuracy and recall in recommendations. This research attests that the contrastive learning framework provides an effective strategy for addressing the long-tail recommendation problem. © 2024 IEEE.

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：