Data privacy has become an increasingly important issue in Machine Learning (ML) , where many approaches have been developed to tackle this challenge, e.g., cryptography ( Homomorphic Encryption (HE) Differential Privacy (DP) ) and collaborative training (Secure Multi-Party Computation (MPC) Distributed Learning, Federated (FL) ). These techniques a particular focus on data encryption or secure local computation. They transfer the intermediate information third party compute final result....

Federated Learning (FL) has emerged as a powerful paradigm for training Machine (ML), particularly Deep (DL) models on multiple devices or servers while maintaining data localized at owners' sites. Without centralizing data, FL holds promise scenarios where integrity, privacy and security are critical. However, this decentralized process also opens up new avenues opponents to launch unique attacks, it been becoming an urgent need understand the vulnerabilities corresponding defense...

To predict future trends based on the data from sensors is an important technology for many applications, such as Internet of Things, smart cities, etc. Based predicted results, further decisions and system controls can be made. Raw sensor sets are often complex non-linear with noise, which results in difficulty accurate prediction. This paper proposes a distributed deep prediction network covariance intersection (CI) fusion algorithm learning networks, long-term short-term memory networks...

Conventional machine learning methodologies require the centralization of data for model training, which may be infeasible in situations where sharing limitations are imposed due to concerns such as privacy and gradient protection. The Federated Learning (FL) framework enables collaborative a shared without necessitating or among proprietors. Nonetheless, this paper, we demonstrate that generalization capability joint is suboptimal Non-Independent Non-Identically Distributed (Non-IID) data,...

Federated Learning (FL) is a widely adopted privacy-preserving machine learning approach where private data remains local, enabling secure computations and the exchange of local model gradients between clients third-party parameter servers. However, recent findings reveal that privacy may be compromised sensitive information potentially recovered from shared gradients. In this study, we offer detailed analysis novel perspective on understanding gradient leakage problem. These theoretical...

The development of the Internet Things (IoT) stimulates many research works related to Multimedia Communication Systems (MCS), such as human face detection and tracking. This trend drives numerous progressive methods. Among these methods, deep learning‐based methods can spot patch in an image effectively accurately. Many people consider tracking detection, but they are two different techniques. Face focuses on a single image, whose shortcoming is obvious, unstable unsmooth position when...

Typical machine learning approaches require centralized data for model training, which may not be possible where restrictions on sharing are in place due to, instance, privacy and gradient protection. The recently proposed Federated Learning (FL) framework allows a shared collaboratively without being or among owners. However, we show this paper that the generalization ability of joint is poor Non-Independent Non-Identically Distributed (Non-IID) data, particularly when Averaging (FedAvg)...

Vision-Language Models (VLMs) play a crucial role in robotic manipulation by enabling robots to understand and interpret the visual properties of objects their surroundings, allowing them perform based on this multimodal understanding. However, understanding object attributes spatial relationships is non-trivial task but critical tasks. In work, we present new dataset focused attribute assignment novel method utilize VLMs with task-oriented, high-level input. dataset, between are manually...

Unmanned Aerial Vehicle (UAV) swarms are increasingly deployed in dynamic, data-rich environments for applications such as environmental monitoring and surveillance. These scenarios demand efficient data processing while maintaining privacy security, making Federated Learning (FL) a promising solution. FL allows UAVs to collaboratively train global models without sharing raw data, but challenges arise due the non-Independent Identically Distributed (non-IID) nature of collected by UAVs. In...

Reliable object grasping is one of the fundamental tasks in robotics. However, determining pose based on single-image input has long been a challenge due to limited visual information and complexity real-world objects. In this paper, we propose Triplane Grasping, fast decision-making method that relies solely single RGB-only image as input. Grasping creates hybrid Triplane-Gaussian 3D representation through point decoder triplane decoder, which produce an efficient high-quality...

Federated learning (FL) is a powerful Machine Learning (ML) paradigm that enables distributed clients to collaboratively learn shared global model while keeping the data on original device, thereby preserving privacy. A central challenge in FL effective aggregation of local weights from disparate and potentially unbalanced participating clients. Existing methods often treat each client indiscriminately, applying single proportion entire model. However, it empirically advantageous for weight...

Federated Learning (FL) has emerged as a powerful paradigm for training Machine (ML), particularly Deep (DL) models on multiple devices or servers while maintaining data localized at owners’ sites. Without centralizing data, FL holds promise scenarios where integrity, privacy and security are critical. However, this decentralized process also opens up new avenues opponents to launch unique attacks, it been becoming an urgent need understand the vulnerabilities corresponding defense...

This review paper takes a comprehensive look at malicious attacks against FL, categorizing them from new perspectives on attack origins and targets, providing insights into their methodology impact. In this survey, we focus threat models targeting the learning process of FL systems. Based source target attack, categorize existing four types, Data to Model (D2M), (M2D), (M2M) composite attacks. For each type, discuss defense strategies proposed, highlighting effectiveness, assumptions...

AbstractAbstractCorrosion has been one of the most serious safety problems in aviation, and costly maintenance problem. Contributing to problematic nature corrosion is complexity its mechanisms randomness with which it occurs. The reliability a structure affected by normally based on forms erosion severity. Existing research this area primary focus induced failure due inadequate residual structural strength. However, valuable recognise that some product failures are not critically dependent...