检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on artificial intelligence, AAAI 2025

作者： Li, Linhao Tan, Yongzhang Yang, Siyuan Cheng, Hao Dong, Yongfeng Yang, Liang School of Artificial Intelligence and Data Science Hebei University of Technology China College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (纸本)157735897X

Few-Shot Class-Incremental Learning (FSCIL) aims to continuously learn new classes from a limited set of training samples without forgetting knowledge of previously learned classes. Conventional FSCIL methods typically build a robust feature extractor during the base training session with abundant training samples and subsequently freeze this extractor, only fine-tuning the classifier in subsequent incremental phases. However, current strategies primarily focus on preventing catastrophic forgetting, considering only the relationship between novel and base classes, without paying attention to the specific decision spaces of each class. To address this challenge, we propose a plug-and-play Adaptive Decision Boundary Strategy (ADBS), which is compatible with most FSCIL methods. Specifically, we assign a specific decision boundary to each class and adaptively adjust these boundaries during training to optimally refine the decision spaces for the classes in each session. Furthermore, to amplify the distinctiveness between classes, we employ a novel inter-class constraint loss that optimizes the decision boundaries and prototypes for each class. Extensive experiments on three benchmarks, namely CIFAR100, miniImageNet, and CUB200, demonstrate that incorporating our ADBS method with existing FSCIL techniques significantly improves performance, achieving overall state-of-the-art results. Copyright © 2025, Association for the Advancement of artificial intelligence (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A robust inlier identification algorithm for point cloud registration via 0-minimization 38

A robust inlier identification algorithm for point cloud reg...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Jiang, Yinuo Tang, Xiuchuan Cheng, Cheng Yuan, Ye School of Artificial Intelligence and Automation Huazhong University of Science and Technology China Department of Automation Tsinghua University China

Correspondences in point cloud registration are prone to outliers, significantly reducing registration accuracy and highlighting the need for precise inlier identification. In this paper, we propose a robust inlier identification algorithm for point cloud registration by reformulating the conventional registration problem as an alignment error 0-minimization problem. The 0-minimization problem is formulated for each local set, where those local sets are built on a compatibility graph of input correspondences. To resolve the 0-minimization, we develop a novel two-stage decoupling strategy, which first decouples the alignment error into a rotation fitting error and a translation fitting error. Second, null-space matrices are employed to decouple inlier identification from the estimation of rotation and translation respectively, thereby applying Bayes Theorem to 0-minimization problems and solving for fitting errors. Correspondences with the smallest errors are identified as inliers to generate a transformation hypothesis for each local set. The best hypothesis is selected to perform registration. We demonstrate that the proposed inlier identification algorithm is robust under high outlier ratios and noise through experiments. Extensive results on the KITTI, 3DMatch, and 3DLoMatch datasets demonstrate that our method achieves state-of-the-art performance compared to both traditional and learning-based methods in various indoor and outdoor scenes. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multiscale Complementary Semantic Enhancement Network for Welding Seam X-ray Image Defect Detection 7

Multiscale Complementary Semantic Enhancement Network for We...

引用

7th International Symposium on Autonomous Systems, ISAS 2024

作者： Liu, Weipeng Wang, Yuheng Chen, Haiyong Hebei University of Technology School of Artificial Intelligence and Data Science Tianjin300132 China

ISBN: (纸本)9798350363173

The core challenge in the analysis of welding seam X-ray images is the problem of multiscale defects, where defects can range in size and shape from tiny pores to larger unpenetrated areas. However, the current methods have limitations, such as information loss during feature extraction and inadequate scale adjustment strategies during feature fusion, making it difficult for models to capture complex semantic information and retain key details effectively. In this paper, we propose a Multiscale Complementary Semantic Enhancement Network (MCSEN), which includes a Multi-level Semantic Enhancement (MSE) module and a Multi-level Feature Complementary Aggregation (MFCA) module. The MSE module improves multiscale feature learning through graph convolution-guided feature aggregation, effectively mitigating the problem of information loss and enhancing features' distinguishability and semantic richness. The MFCA module resolves the issues of scale inconsistency and semantic discontinuity in feature fusion by adaptively generating fusion weights and finely adjusting feature allocation. Comprehensive experiments show that MCSEN achieves an overall average precision of 91.9%, significantly higher than other models. © 2024 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Simultaneous Learning Intensity and Optical Flow From High-Speed Spike Stream

引用

IEEE Transactions on Circuits and Systems for Video technology 2025年第6期35卷 5126-5139页

作者： Zhu, Lin Yan, Weiquan Chang, Yi Tian, Yonghong Huang, Hua Beijing Institute of Technology School of Computer Science and Technology Beijing100081 China Peng Cheng Laboratory Shenzhen518055 China Huazhong University of Science and Technology School of Artificial Intelligence and Automation Wuhan430074 China Peking University School of Computer Science Beijing100081 China Beijing Normal University School of Artificial Intelligence Beijing100875 China

Bio-inspired vision sensors, which emulate the human retina by recording light intensity as binary spikes, have gained increasing interest in recent years. Among them, the spike camera is capable of perceiving fine textures by simulating a small retinal region called the fovea and producing high temporal resolution (20,000 Hz) spatiotemporal spike streams. To bridge the gap between binary spike streams and human vision in high-speed scenes, reconstructing intensity and optical flow from high temporal resolution spikes is particularly important. In this paper, we present a hybrid SNN-ANN network designed for simultaneous intensity and optical flow learning from spike streams. To adaptively extract spatial and temporal features from continuous spike streams, we propose a spiking neuron module with dense connections that efficiently processes both short-term and long-term spike data, while maintaining low power consumption characteristics. Subsequently, we introduce two decoders for optical flow and intensity estimation that complement each other. A temporal-aware warping module, based on flow features, is specifically designed to align the temporal features of the intensity decoder, thereby reducing motion artifacts. Concurrently, improved intensity features contribute to more accurate flow feature predictions, resulting in a mutually beneficial relationship within our network. To evaluate the effectiveness of our proposed network, we conduct experiments on both simulated and real spike datasets. Our network outperforms existing state-of-the-art spike-based reconstruction and optical flow estimation methods, demonstrating its potential for advancing the field of bio-inspired vision sensors. © 1991-2012 IEEE.

关键词： Optical flows

来源：评论

学校读者我要写书评

暂无评论

Analysis and Prediction of Polycystic Ovarian Syndrome Using ML Classifiers 1

引用

8th International Conference on Emerging Research in Computing, Information, Communication and Applications, ERCICA 2023

作者： Sanjay, Lavanya Nayak, Neha Sriram, Disha Kodipalli, Ashwini Rao, Trupthi Kamal, Shoaib Department of Artificial Intelligence and Data Science Global Academy of Technology Karnataka Bangalore India

ISBN: (数字)9789819976331

ISBN: (纸本)9789819976324

Polycystic Ovarian Syndrome (PCOS) is an endocrine disorder which is common in females of reproductive age. A large number of females of reproductive age are affected by PCOS due to adverse change in their lifestyle. As of the current trend between 2.2 and 26.7%, women of childbearing age have PCOS. Visiting a gynaecologist might be a taboo for many women, let alone the cost of running multiple tests. So, using medical expertise and feature selection we have filtered and obtained basic parameters that can predict if a woman has PCOS or not with minimal number of medical tests, thus reducing cost and time. In this study, we have used feature selection and classification algorithms to detect if a woman has PCOS or not. The algorithms used are Logistic Regression and Random Forest of which we have concluded that Random Forest provides a higher accuracy. © 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Logistic regression

来源：评论

学校读者我要写书评

暂无评论

Analyzing and predicting the performance of players using machine learning 4

Analyzing and predicting the performance of players using ma...

引用

4th International Conference on Power, Energy, Control and Transmission Systems, ICPECTS 2024

作者： Uma Mageswari, T. Pavithra, B. Priyashree, S.S. Sharu Mithra, S. Sri Sairam Institute of Technology Department of Artificial Intelligence and Data Science Chennai India

ISBN: (数字)9798331508845

ISBN: (纸本)9798331508845

The ensemble method employs a combination of logistic regression and random forest classification to analyze cricket player performance. Using datasets from the Indian Premier League (IPL) and India T20 series, key performance metrics such as batting average and strike rate are examined. This strategic approach enhances the precision of player assessment, providing insights into individual proficiency as well as broader implications for team composition and match strategies. The research highlights the critical role of logistic regression and random forest classification in advancing cricket analytics. © 2024 IEEE.

关键词： Logistic regression

来源：评论

学校读者我要写书评

暂无评论

Proliferative Diabetic Retinopathy Detection and Progression Analysis 4

Proliferative Diabetic Retinopathy Detection and Progression...

引用

4th International Conference on Soft Computing for Security Applications, ICSCSA 2024

作者： Rathinasamy, Dhivya Logesh, S. Shri Guru Varshini, S. Senthilkumari, P. PSNA College of Engineering and Technology Department of Artificial Intelligence and Data Science Dindigul India

ISBN: (纸本)9798331515720

Proliferative Diabetic Retinopathy (PDR) is a severe complication of diabetes that can lead to vision loss if not detected and managed promptly. This data analysis project aims to develop an advanced diagnostic system for PDR detection using retinal images. Going beyond conventional approaches, proposed a novel methodology that integrates multi-modal data and employs explainable artificial intelligence (AI) techniques to enhance accuracy and interpretability. The project utilizes a diverse dataset, including retinal images, patient histories, genetics, and lifestyle data. A multi-modal fusion framework extracts features capturing subtle Proliferative Diabetic Retinopathy (PDR) variations. Alongside precise diagnosis, the study introduces longitudinal analysis for PDR progression *** model not only detects PDR but also provides transparent insights into its decision-making process, empowering clinicians to validate and understand the predictions. The proposed design involves using a Convolutional Neural Network (CNN) to classify images of the retina into different stages of Diabetic Retinopathy. The model will be evaluated using the Confusion Matrix and True Value and Predicted Value Comparison. The outcomes of this project extend beyond conventional PDR detection by offering an interpretable AI-powered tool for clinicians. The system holds promise for revolutionizing PDR management, promoting early detection, and providing tailored interventions for patients. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Camellia Sinensis (Tea) Plant Disease Classification Using RESNET

Camellia Sinensis (Tea) Plant Disease Classification Using R...

引用

2024 International Conference on science, technology, Engineering and Management, ICSTEM 2024

作者： Deepa, C. Sanjay, S.M. Mosses, Victor Yugeshwaran, M. Sam Jebaraj, A. Department of Artificial Intelligence and Data Science KIT- Kalaignarkarunanidhi Institute of Technology Coimbatore India

ISBN: (纸本)9798350376913

Tea, a globally consumed beverage, holds a significant place in the agricultural sector. However, the cultivation of tea is threatened by various diseases that can adversely affect the quality and yield of the tea crop. Timely detection and management of these diseases are essential for sustainable tea production. In this research, we present a novel methodology to address this challenge by leveraging deep learning techniques, specifically the Residual Neural Network (ResNet), for the automated identification of tea plant diseases. The primary objective of this research is to develop a robust and accurate system that can assist farmers and researchers in identifying diseases in tea plants. Traditional methods of disease identification often require expert knowledge and can be time-consuming. In contrast, our proposed approach harnesses the power of computer vision and machine learning to streamline the identification process, enabling faster and more accurate diagnoses. To accomplish this, we have curated a diverse and comprehensive dataset of tea plant images, capturing both healthy plants and various diseases, including but not limited to blister blight, gray blight, and root rot. The quality an,d diversity of the dataset play a pivotal role in training a reliable disease identification model. We employ a ResNet architecture, a state-of-the-art convolutional neural network known for its effectiveness in image classification tasks. © 2024 IEEE.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

DAO to HANOI via DeSci: AI Paradigm Shifts from AlphaGo to ChatGPT

引用

IEEE/CAA Journal of Automatica Sinica 2023年第4期10卷 877-897页

作者： Qinghai Miao Wenbo Zheng Yisheng Lv Min Huang Wenwen Ding Fei-Yue Wang School of Artificial Intelligence University of Chinese Academy of SciencesBeijing 100049China School of Computer Science and Artificial Intelligence Wuhan University of TechnologyWuhan 430070China Institute of Automation Chinese Academy of SciencesBeijing 100190China Institute of System Engineering Macao University of Science and TechnologyMacaoChina

From AlphaGo to ChatGPT,the field of AI has launched a series of remarkable achievements in recent ***,comparing,and summarizing these achievements at the paradigm level is important for future AI innovation,but has not received sufficient *** this paper,we give an overview and perspective on machine learning ***,we propose a paradigm taxonomy with three levels and seven dimensions from a knowledge ***,we give an overview on three basic and twelve extended learning paradigms,such as Ensemble Learning,Transfer Learning,etc.,with figures in unified *** further analyze three advanced paradigms,i.e.,AlphaGo,AlphaFold and ***,to enable more efficient and effective scientific discovery,we propose to build a new ecosystem that drives AI paradigm shifts through the decentralized science(DeSci)movement based on decentralized autonomous organization(DAO).To this end,we design the Hanoi framework,which integrates human factors,parallel intelligence based on a combination of artificial systems and the natural world,and the DAO to inspire AI innovations.

关键词： ChatGPT decentralized science(DeSci) decentralized autonomous organization(DAO) machine learning paradigm shift

来源：评论

学校读者我要写书评

暂无评论

Day and Night City View Image Translations Using Cycle GAN 6

Day and Night City View Image Translations Using Cycle GAN

引用

6th IEEE International Conference on Recent Advances in Intelligent Computational Systems, RAICS 2024

作者： Somanna, A.G. Sidduprasad, M. Kodipalli, Ashwini Rao, Trupthi Global Academy of Technology Department of Artificial Intelligence and Data Science Karnataka Bengaluru India

ISBN: (纸本)9798350381689

CycleGAN, a type of Generative Adversarial Networks (GANs), is used in this study to tackle the difficult problem of translating cityscape photographs between day and night settings. Day-to-night picture translation is changing well-lit daytime images into nocturnal landscapes using artificial lighting, unique color palettes, and atmospheric effects, and vice versa. Previously, this work required labour-intensive manual involvement or specialised picture algorithms, but deep learning and GANs have introduced automation and better quality to this industry. CycleGAN excels at unpaired picture translation, even when there are no obvious correspondences between daylight and nighttime datasets. It makes use of the idea of cycle consistency to ensure proper translation between picture domains. This study focuses on training a CycleGAN model and found encouraging results. Beyond technical aspects, this paper highlights practical applications spanning film production, gaming, urban planning, traffic management, photography, environmental impact assessment, and more. CycleGAN's adaptability offers opportunities to enhance realism, aesthetics, and decision-making across various domains. In summary, this research underscores the potential of CycleGAN in day-to-night image translation, unlocking new possibilities in visualizing urban environments and influencing diverse industries and applications. © 2024 IEEE.

关键词： Decision making

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：