检索结果-内蒙古大学图书馆

2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion

作者： Sun, Ning Zhu, Shuwei School of Artificial Intelligence and Computer Science Jiangnan University Wuxi China

ISBN: (纸本)9798400704956

The Make-up Air Unit (MAU) is an air conditioning device serving semiconductor cleanrooms, which provides constant temperature and humidity for fresh air. The common air handling sections of the MAU can be divided into: fresh air filtration, preheating, precooling, water-washing humidification, recooling, reheating, air supply, and high-efficiency filtration. The commonly used PID control methods for MAU have limitations, this work proposes a proactive energy-saving optimization control method based on machine learning and genetic algorithm (GA). Firstly, the machine learning methods are used to train historical data of MAU, resulting in a data-driven prediction model of energy consumption for the system. Subsequently, GA is used to optimize energy in cold and hot water systems. It facilitates the dynamic adjustment of the regulating valve opening for the cold and hot water coil in the fresh air unit, responding to real-time variations in outdoor air conditions. Meanwhile, it ensures that the supply air temperature and humidification adhere to specified requirements, thereby reducing the energy consumption associated with cold and hot water usage in the MAU. © 2024 Copyright held by the owner/author(s).

关键词： Preheating

来源：评论

学校读者我要写书评

暂无评论

Deep learning-based semantic segmentation of human features in bath scrubbing robots

引用

Biomimetic intelligence & Robotics 2024年第1期4卷 70-79页

作者： Chao Zhuang Tianyi Ma Bokai Xuan Cheng Chang Baichuan An Minghuan Yin Hao Sun School of Artificial Intelligence and Data Science Hebei University of TechnologyTianjin 300401China Engineering Research Center of Intelligent Rehabilitation Device and Detection Technology Ministry of EducationTianjin 300401China

With the rise in the aging population,an increase in the number of semidisabled elderly individuals has been noted,leading to notable challenges in medical and healthcare,exacerbated by a shortage of nursing *** study aims to enhance the human feature recognition capabilities of bath scrubbing robots operating in a water fog *** investigation focuses on semantic segmentation of human features using deep learning ***,3D point cloud data of human bodies with varying sizes are gathered through light detection and ranging to establish human ***,a hybrid filtering algorithm was employed to address the impact of the water fog environment on the modeling and extraction of human ***,the network is refined by integrating the spatial feature extraction module and the channel attention module based on *** results indicate that the algorithm adeptly identifies feature information for 3D human models of diverse body sizes,achieving an overall accuracy of 95.7%.This represents a 4.5%improvement compared with the PointNet network and a 2.5%enhancement over mean intersection over *** conclusion,this study substantially augments the human feature segmentation capabilities,facilitating effective collaboration with bath scrubbing robots for caregiving tasks,thereby possessing significant engineering application value.

关键词： 3D point cloud Human model LiDAR Semantic segmentation

来源：评论

学校读者我要写书评

暂无评论

A review of generative models for virtual human motion driving

引用

science China(Information sciences) 2025年

作者： Pengcheng WANG Haolun PENG Siyang MAO Shaojiang WANG Jie TANG Department of Computer Science and Technology Tsinghua University School of Artificial Intelligence Nanjing University Nanjing Institute of Software Technology School of Computer Science and Technology Beijing Institute of Technology Institute of Software Chinese Academy of Sciences Institute of Industrial Artificial Intelligence Chinese Academy of Sciences

Virtual human motion driving focuses on generating and controlling realistic human motions, from facial expressions to body movements. These motions are driven by various types of input signals, such as visual and acoustic features,textual prompts, or a combination thereof. This survey delivers an in-depth examination of generative models for virtual human motion driving, with a specific emphasis on recent models. A taxonomy of virtual human motion driving networks designed for talking-face and human-pose generation is provided. The former mainly concentrates on lip synchronization,differentiation of emotions, and personalized expressions, while the latter mainly includes co-speech gesture generation and text-to-motion prediction. Moreover, available datasets and evaluation metrics for virtual human motion driving tasks are discussed, applications and real products related to virtual human motion driving are explored, along with their challenges,limitations, and potential future developments. The objective of this survey is to gain a comprehensive understanding of the present advancements in talking-face and human-pose generation models, with a focus on the future potential of virtual human motion driving. This endeavor aims to lay the groundwork for the development of extensive applications for virtual humans.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep learning-based clustering method for single-cell RNA data 7

Deep learning-based clustering method for single-cell RNA da...

引用

7th International Symposium on Autonomous Systems, ISAS 2024

作者： Li, Yuxuan Chen, Xin Zhu, Ruishu Wuhan University of Technology School of Science Wuhan China Wuhan University of Technology School of Computer and Artificial Intelligence Wuhan China

ISBN: (纸本)9798350363173

This study proposes innovative ideas in feature engineering for the clustering problem of single-cell RNA data and provides new technical paths for existing clustering methods. Through the proposed feature engineering techniques, we effectively address the curse of dimensionality in the clustering analysis of single-cell RNA data. We introduce an Auto Encoder network that facilitates the reduction and reconstruction of data dimensionality. Meanwhile, this project proposes a novel hidden layer selection method and further utilizes the ISOMAP technique to achieve deeper dimensionality reduction. This study compares two clustering technology paths: one based on cell data and the other on cell feature vectors. It achieves innovation in clustering technology paths by introducing the OPTICS algorithm as a supplement to the k-means algorithm, introducing UMAP and t-SNE for dimensionality reduction, and by exploring the MDS algorithm for low-dimensional data embedding. By introducing internal and external clustering performance evaluation metrics, this thesis verifies that the technical combination of the proposed MDS dimensionality reduction algorithm combined with the k-means method constitutes an innovative clustering technology path that outperforms the traditional methods, and promotes the innovation and improvement of clustering analysis of single-cell RNA data. © 2024 IEEE.

关键词： RNA

来源：评论

学校读者我要写书评

暂无评论

LGNet: Linear Graph Representation for Efficient Cold-Start Recommendations

LGNet: Linear Graph Representation for Efficient Cold-Start ...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Li, Zhenglong Luo, Ruiqi Wang, Bangchao Li, Lin Zhong, Xian School of Computer Science and Artificial Intelligence Wuhan Textile University China Hubei Key Laboratory of Transportation Internet of Things School of Computer Science and Artificial Intelligence Wuhan University of Technology China

ISBN: (纸本)9798350368741

Graph Convolutional Networks (GCNs) demonstrate significant potential in recommendation systems but face difficulties with the cold-start problem, especially in integrating new nodes during inference. The typical solution leverages meta-learning for few-shot learning, though it often fails to fully capture collaborative filtering between nodes. In this paper, we revisit the node embedding propagation algorithm in GCNs, emphasizing the importance of collaborative filtering and elucidating the relation between high-order and low-order embeddings. Given the substantial interaction data required for training recommendation models, maintaining a simple model structure remains crucial. To address these challenges, we propose the Linear Graph Network (LGNet), which theoretically compresses multi-layer GCNs into a single layer, enabling the embedding of new nodes during inference. Experimental results on benchmark datasets for link prediction and user cold-start tasks demonstrate that LGNet outperforms existing methods. The code will be available at https://***/kunbeibei/LGNet. © 2025 IEEE.

关键词： Cold-Start Problem Graph Convolutional Networks Linear Representation Recommendation Systems

来源：评论

学校读者我要写书评

暂无评论

Research on PM2.5 Concentration Prediction Model Based on RNCL-A 5

Research on PM2.5 Concentration Prediction Model Based on RN...

引用

5th International Conference on Advances in Electrical Engineering and computer Applications, AEECA 2024

作者： Zhang, Xujuan Jia, Wenhao Sun, Lijuan Li, Lan Zhu, Changyu Lanzhou Institute of Technology School of Computer Science and Artificial Intelligence Lanzhou China Hefei University of Technology School of Computer and Information Lanzhou China

ISBN: (纸本)9798350355253

As industrialization and urbanization progress at a swift pace, the issue of air pollution has intensified significantly, with PM2.5, a pivotal constituent of smog, emerging as a grave hazard to human wellbeing, the ecological landscape, and economic prosperity alike. Therefore, it is particularly important to make scientific and accurate prediction of PM2.5 concentration. In order to improve the accuracy and robustness of PM2.5 concentration prediction, this paper constructs an RNCL-A integrated model aiming at the shortcomings of existing prediction models in generalization ability and stability. Taking the publicly available Beijing PM dataset as an example, the Stacking integration framework is used, combining the advantages of ResNet+Attention, CNN, and LSTM. Among them, ResNet+Attention model is used to extract deep-level features and important information in data, CNN model focuses on capturing spatial features, and LSTM model is responsible for processing long-term dependencies in time series data to improve the accuracy of prediction. Through case study, and evaluation and comparison with ResNet-LSTM, CNN-LSTM and LSTM models, the results show that RNCL-A is superior to other models in multiple evaluation indicators, and this model can not only effectively capture the spatiotemporal characteristics and nonlinear relations in the data, but also effectively capture the spatial and temporal characteristics of the data. It can also improve the accuracy and stability of prediction through multi-model fusion, which proves its application potential in the field of air quality prediction. This study not only provides a new deep learning solution for PM2.5 concentration prediction, but also provides a reference for other environmental monitoring problems. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Real-Time Efficient Environment Compression and Sharing for Multi-Robot Cooperative Systems

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-14页

作者： Zheng, Lanxiang Xu, Kai Jiang, Jinqi Wei, Mingxin Zhou, Boyu Cheng, Hui School of Computer Science and Engineering Sun Yat-Sen University Guangzhou China School of Systems Science and Engineering Sun Yat-Sen University Guangzhou China School of Future Technology Harbin Institute of Technology Harbin China School of Artificial Intelligence Sun Yat-Sen University Zhuhai China

Efficient environment sharing is crucial for multi-robot tasks, such as exploration and navigation. However, real-time environment sharing faces significant challenges due to limited communication bandwidth. Inspired by image JPEG compression, this paper presents a novel solution for efficient compression and real-time sharing of environmental point clouds. The framework directly maps the 3D point cloud obtained by the sensor into a panorama, implicitly reducing data dimensions. An event-trigger mechanism, based on the visibility of the point cloud, selectively merges consecutive frames of the point cloud into same panorama, thereby avoiding the transmission of redundant data. To reduce the proportion of invalid data in the panorama, a Mixed Integer Programming (MIP) problem is formulated to extract valuable segments from the panorama before compression. The point cloud is then compressed in the frequency domain, achieving high compression performance and decompression accuracy. The lightweight architecture without GPU acceleration makes the framework easily deployable and suitable for real-time environment sharing in multi-robot systems. Simulations and real-world experiments in various scenarios validate the effectiveness of the proposed method. IEEE

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

A joint improved DETR network face recognition algorithm based on TSM module and DANN. 23

A joint improved DETR network face recognition algorithm bas...

引用

4th International Conference on Computing, Networks and Internet of Things, CNIOT 2023

作者： Ye, Ziheng Wuhan University of Technology School of Computer Science and Artificial Intelligence China

ISBN: (纸本)9798400700705

As one of the most mature application fields of artificial intelligence, face recognition system has been widely used in production and life. But at the same time as large-scale commercialization, face recognition technology also faces more challenges. How to further improve the accuracy of face recognition function and improve the defense ability of the recognition system against face against samples is an important research direction of algorithms. At present, the algorithm mostly focuses on the picture features of a single portrait, ignores the details difference in the time domain of the fake video, and has the problems of weak generalization ability and overfitting of the model. In this paper, an improved DETR network is proposed, which uses the TSM module to perform time domain displacement on the extracted video features and averages the time features of pooled learning samples to better distinguish dynamic adversarial sample instances. At the same time, DANN is introduced as a systematic classification and discrimination network, which uses the domain classifier to domain discrimination of the feature space and uses the adversarial loss function to update the feature extractor and domain classifier parameters. Actual tests show that the recognition accuracy of the network on the FaceForensics dataset is improved by an average of 4%-7% compared with the improvement period, the error rate is less than 7%, and the model recognition speed index is not higher 400ms, which proves that the model has high accuracy rate and good real-Time solving ability. © 2023 ACM.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Subgraph-based Self-Supervised Learning Framework for Enzymatic Reaction Feasibility Prediction

Subgraph-based Self-Supervised Learning Framework for Enzyma...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Yang, Feng Liu, Juan Zhang, Qiang Yang, Zhihui Liu, Jianghang Wu, Guangsheng Wuhan University Institute of Artificial Intelligence School of Computer Science China Xinyu University School of Mathematics and Computer Science China

ISBN: (纸本)9798350386226

Enzymatic Reaction feasibility prediction is used to determine whether the reaction generated by computational methods can actually occur, which can effectively reduce the complexity of synthetic pathway design. Existing methods often use SMILES or molecular fingerprints to represent molecules, resulting in a lack of molecular structural information. Although some GNNs have been leveraged to solve this problem, the complexity of the intra and inter-substructures interactions of molecules in microorganisms makes it difficult for traditional GNNs to accurately model it. To address these problems, we propose a subgraph-based self-supervised learning framework to predict the feasibility of enzymatic reactions. Specifically, we first propose a subgraph-based two-branch graph neural network. This network leverages the atom graph and substructure graph of a molecule to thoroughly capture its structural and semantic information. Besides, a subgraph interaction module is designed to facilitate the full integration of features. Subsequently, we propose a domain knowledge-guided self-supervised learning task, utilizing molecular fingerprints and substructures to capture the consistency between them effectively. The experimental results show that the proposed method outperforms existing state-ofthe-art methods significantly on all datasets. © 2024 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Detection of Tampered Images with Convolutional Neural Networks (VGG-16) 1st

Detection of Tampered Images with Convolutional Neural Netwo...

引用

1st International Conference on data Engineering and Machine intelligence, ICDEMI 2023

作者： Asiya Afreen Rohith, Gandam Goud, Guluri Manitha Tharun, Vallabhu Abhiram, Kukatla School of Computer Science & Artificial Intelligence SR University Telangana Warangal India Computer Science & Artificial Intelligence SR University Telangana Warangal India

ISBN: (纸本)9789819776153

In this project, we centered around the robust detection of tampered images through the implementation of the VGG-16 algorithm, a Convolutional Neural Network (CNN) known for its effectiveness in image analysis. We curated a diverse dataset from Kaggle, incorporating additional custom images, and meticulously divided it into distinct training and testing sets. Leveraging the capabilities of the VGG-16 model, our system underwent rigorous training and evaluation, ultimately achieving an impressive result. This outcome establishes our project as a powerful forgery detection tool. the user interface, developed with Tkinter, offers a straightforward experience, enabling users to effortlessly upload images and receive prompt classification results clearly indicating whether an image is deemed "Authen-tic" or "Forged." Through this project, we underscore the practical integration of deep learning methodologies into everyday scenarios, providing a user-friendly solution for forgery detection that is accessible to all. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolutional neural networks

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：