Federated recommender systems(FedRecs) have garnered increasing attention recently, thanks to their privacypreserving benefits. However, the decentralized and open characteristics of current FedRecs present at least t...
详细信息
Federated recommender systems(FedRecs) have garnered increasing attention recently, thanks to their privacypreserving benefits. However, the decentralized and open characteristics of current FedRecs present at least two ***, the performance of FedRecs is compromised due to highly sparse on-device data for each client. Second, the system's robustness is undermined by the vulnerability to model poisoning attacks launched by malicious users. In this paper, we introduce a novel contrastive learning framework designed to fully leverage the client's sparse data through embedding augmentation, referred to as CL4FedRec. Unlike previous contrastive learning approaches in FedRecs that necessitate clients to share their private parameters, our CL4FedRec aligns with the basic FedRec learning protocol, ensuring compatibility with most existing FedRec implementations. We then evaluate the robustness of FedRecs equipped with CL4FedRec by subjecting it to several state-of-the-art model poisoning attacks. Surprisingly, our observations reveal that contrastive learning tends to exacerbate the vulnerability of FedRecs to these attacks. This is attributed to the enhanced embedding uniformity, making the polluted target item embedding easily proximate to popular items. Based on this insight, we propose an enhanced and robust version of CL4FedRec(rCL4FedRec) by introducing a regularizer to maintain the distance among item embeddings with different popularity levels. Extensive experiments conducted on four commonly used recommendation datasets demonstrate that rCL4FedRec significantly enhances both the model's performance and the robustness of FedRecs.
The factory has adopted an extensive ecosystem of connected devices and IoT sensors, utilizing cloud computing for real-time decision-making. Secure cloud storage serves as the backbone, managing vast datasets and ena...
详细信息
The factory has adopted an extensive ecosystem of connected devices and IoT sensors, utilizing cloud computing for real-time decision-making. Secure cloud storage serves as the backbone, managing vast datasets and enabling centralized control. By leveraging advanced analytics and machine learning on the cloud, the factory has implemented predictive maintenance, minimizing downtime and optimizing production. The integration of Hybrid PSO-GA for machines and supply chain processes streamlines operations, allowing for remote monitoring and control to enhance operational agility. Cutting-edge advancements in New Generation information Technologies (New IT) are crucial in driving the evolution of smart manufacturing. The proliferation of Internet-connected devices in these environments generates substantial data throughout the product lifecycle. Adopting a cloud-based smart manufacturing strategy provides numerous services and applications for analysing massive datasets and fostering significant cooperation in manufacturing operations. However, challenges such as latency, bandwidth congestion, and network unavailability hinder its effectiveness for real-time applications requiring fast, low-latency performance. These issues are efficiently addressed by integrating cloud computing with edge computing, extending the cloud’s capabilities to the edge. This paper presents a hierarchical reference architecture for smart manufacturing, leveraging cloud computing. The proposed approach employs a hybrid PSO-GA scheduling function that combines Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) to optimize task start times and reduce latency. The optimal solution from this hybrid approach updates task start times, with subsequent scheduling performed using a selected algorithm. The proposed novel hybrid PSO-GA model integrates AI-driven optimization, IoT, and digital twins to enhance real-time decision-making and adapt to dynamic data streams in smart manufacturing. Its
Finding an appropriate subset of agents (a team) from a larger pool of agents (the source set) so that the team exhibits a desired quality is the essence of the team formation problem. This problem is recognized to ha...
详细信息
The life expectancy of a population is a vital measure of its overall health and healthcare quality. This study use machine learning methods, notably XGBoost, to predict life expectancy in industrialized and emerging ...
详细信息
In serverless computing, the service provider takes full responsibility for function management. However, serverless computing has many challenges regarding data security and function scheduling. To address these chal...
详细信息
In scenarios such as vehicle radiation monitoring and unmanned aerial vehicle radiation detection,rapid measurements using a NaI(Tl)detector often result in low photon counts,weak characteristic peaks,and significant ...
详细信息
In scenarios such as vehicle radiation monitoring and unmanned aerial vehicle radiation detection,rapid measurements using a NaI(Tl)detector often result in low photon counts,weak characteristic peaks,and significant statistical *** issues can lead to potential failures in peak-searching-based identification *** address the low precision associated with short-duration measurements of radionuclides,this paper proposes an identification algorithm that leverages heterogeneous spectral transfer to develop a low-count energy spectral identification *** experiments demonstrated that transferring samples from 26 classes of simulated heterogeneous gamma spectra aids in creating a reliable model for measured gamma *** only 10%of target domain samples used for training,the accuracy on real low-count spectral samples was 95.56%.This performance shows a significant improvement over widely employed full-spectrum analysis methods trained on target domain *** proposed method also exhibits strong generalization capabilities,effectively mitigating overfitting issues in low-count energy spectral classification under short-duration measurements.
Accurate segmentation of brain tumor regions in MRI images is essential for monitoring tumor growth. In view of this, several automated brain tumor segmentation models are proposed. U-Net is one of the most popular mo...
详细信息
This paper presents a resilience-driven framework leveraging advanced control technologies, particularly a Markov chain approach, to enhance the robustness of peer-to-peer (P2P) energy trading networks under Low Proba...
详细信息
There is a growing interest in sustainable ecosystem development, which includes methods such as scientific modeling, environmental assessment, and development forecasting and planning. However, due to insufficient su...
详细信息
The area of brain-computer interface research is widely spreading as it has a diverse array of potential applications. Motor imagery classification is a boon to several people with motor impairment. Low accuracy and d...
详细信息
暂无评论