Manual annotation for crowd counting remains labor-intensive and costly. Although existing semi-supervised methods partially alleviate this burden, they still face significant challenges regarding the quality of gener...
详细信息
Manual annotation for crowd counting remains labor-intensive and costly. Although existing semi-supervised methods partially alleviate this burden, they still face significant challenges regarding the quality of generated pseudo-labels and the utilization of unlabeled data. To address these issues, we propose a novel semi-supervised crowd counting framework, called Point-Adaptive Teacher (PAT). This framework integrates Adaptive Soft Threshold (AST) and contrastive learning to enhance pseudo-label quality and effectively leverage unlabeled data. Specifically, we employ the Swin Transformer as the backbone and develop Swin-P2PNet, which captures global contextual information through hierarchical window attention, improving the accuracy of pseudo-labels. Additionally, we design the AST that dynamically adjusts the sample loss weight by combining confidence and uncertainty predictions, thereby alleviating the effect of noise in pseudo-labels. Finally, we introduce a contrastive learning strategy requiring no extra parameters. This strategy enhances the model's ability to learn latent representations from unlabeled data. Extensive experiments have been conducted on three public datasets, namely ShanghaiTech, JHU-Crowd++, and UCF-QNRF. The results demonstrate that our method achieves performance comparable to state-of-the-art methods.
The two-volume set LNAI 6634 and 6635 constitutes the refereed proceedings of the 15th Pacific-Asia Conference on Knowledge Discovery and data Mining, PAKDD 2011, held in Shenzhen, China in May 2011. The total of 32 r...
详细信息
ISBN:
(数字)9783642208416
ISBN:
(纸本)9783642208409
The two-volume set LNAI 6634 and 6635 constitutes the refereed proceedings of the 15th Pacific-Asia Conference on Knowledge Discovery and data Mining, PAKDD 2011, held in Shenzhen, China in May 2011. The total of 32 revised full papers and 58 revised short papers were carefully reviewed and selected from 331 submissions. The papers present new ideas, original research results, and practical development experiences from all KDD-related areas including data mining, machine learning, artificial intelligence and pattern recognition, data warehousing and databases, statistics, knowledge engineering, behavior sciences, visualization, and emerging areas such as social network analysis.
The advancement of mobile multimedia communications, 5G, and Internet of Things (IoT) has led to the widespread use of edge devices, including sensors, smartphones, and wearables. This has generated in a large amount ...
详细信息
The advancement of mobile multimedia communications, 5G, and Internet of Things (IoT) has led to the widespread use of edge devices, including sensors, smartphones, and wearables. This has generated in a large amount of distributed data, leading to new prospects for deep learning. However, this data is confined within data silos and contains sensitive information, making it difficult to be processed in a centralized manner, particularly under stringent data privacy regulations. Federated learning (FL) offers a solution by enabling collaborative learning while ensuring privacy. Nonetheless, data and device heterogeneity complicate FL implementation. This research presents a specialized FL algorithm for heterogeneous edge computing. It integrates a lightweight grouping strategy for homogeneous devices, a scheduling algorithm within groups, and a Split Learning (SL) approach. These contributions enhance model accuracy and training speed, alleviate the burden on resource-constrained devices, and strengthen privacy. Experimental results demonstrate that the GSFL outperforms FedAvg and SplitFed by 6.53× and 1.18×. Under experimental conditions with \(\alpha=0.05\), representing a highly heterogeneous data distribution typical of extreme Non-IID scenarios, GSFL showed better accuracy compared to FedAvg by 10.64%, HACCS by 4.53%, and Cluster-HSFL by 1.16%. GSFL effectively balances privacy protection and computational efficiency for real-world applications in mobile multimedia communications.
This clearly-structured, classroom-tested textbook/reference presents a methodical introduction to the field of CI. Providing an authoritative insight into all that is necessary for the successful application of CI me...
详细信息
ISBN:
(数字)9781447150138
ISBN:
(纸本)9781447158493
This clearly-structured, classroom-tested textbook/reference presents a methodical introduction to the field of CI. Providing an authoritative insight into all that is necessary for the successful application of CI methods, the book describes fundamental concepts and their practical implementations, and explains the theoretical background underpinning proposed solutions to common problems. Only a basic knowledge of mathematics is required. Features: provides electronic supplementary material at an associated website, including module descriptions, lecture slides, exercises with solutions, and software tools; contains numerous examples and definitions throughout the text; presents self-contained discussions on artificial neural networks, evolutionary algorithms, fuzzy systems and Bayesian networks; covers the latest approaches, including ant colony optimization and probabilistic graphical models; written by a team of highly-regarded experts in CI, with extensive experience in both academia and industry.
1 WorkshopTheme data mining as a discipline aims to relate the analysis of large amounts of user data to shed light on key business questions. Web usage mining in particular, a relatively young discipline, investigate...
详细信息
ISBN:
(数字)9783540396635
ISBN:
(纸本)9783540203049
1 WorkshopTheme data mining as a discipline aims to relate the analysis of large amounts of user data to shed light on key business questions. Web usage mining in particular, a relatively young discipline, investigates methodologies and techniques that - dress the unique challenges of discovering insights from Web usage data, aiming toevaluateWebusability,understandtheinterestsandexpectationsofusersand assess the e?ectiveness of content delivery. The maturing and expanding Web presents a key driving force in the rapid growth of electronic commerce and a new channel for content providers. Customized o?ers and content, made possible by discovered knowledge about the customer, are fundamental for the establi- ment of viable e-commerce solutions and sustained and e?ective content delivery in noncommercial domains. Rich Web logs provide companies with data about their online visitors and prospective customers, allowing microsegmentation and personalized interactions. While Web mining as a domain is several years old, the challenges that characterize data analysis in this area continue to be formidable. Though p- processing data routinely takes up a major part of the e?ort in data mining, Web usage data presents further challenges based on the di?culties of assigning data streams to unique users and tracking them over time. New innovations are required to reliably reconstruct sessions, to ascertain similarity and di?erences between sessions, and to be able to segment online users into relevant groups.
暂无评论