Federated learning can collaboratively train AI models while protecting data privacy. In practical industry environment, non-independent and identically distributed (Non-IID) characteristics of affect the effectiveness federated learning. Personalized help resolve this, but it cannot adapt to unknown data. addition, applications also call for trusted training environment remain stable when there are security threats. this article, we propose a credible self-learning (CFSL), based on idea...

This paper proposes a new wireless video communication scheme to achieve high-efficiency transmission over noisy channels. It exploits the idea of model division multiple access (MDMA) and extracts common semantic features across frames. Besides, deep joint source-channel coding (JSCC) is applied overcome distortion caused by The proposed framework collected under name (MDVSC). In particular, temporal relative frames are first transformed into latent space for computing complexity reduction...

Federated learning is a distributed machine approach that achieves collaborative training while protecting data privacy. However, in scenarios, the operational of industrial equipment dynamic and non-independently identically (non-IID). This situation leads to poor performance federated algorithms anomaly detection tasks. Personalized viable solution non-IID problem, but it not effective responding environmental changes. Implementing directed updates model, thereby effectively maintaining...

Federated learning enables multiple clients to learn a general model without sharing local data, and the federated system also improves information security advances responsible artificial intelligence (AI). However, data of different in are non-independently identically distributed (IID), which results weight divergence, especially for complex graph extraction. This article proposes novel feature-contrastive (FcgFed) approach improve robustness data. First, we design an architecture FcgFed...

Federated meta-learning solves many challenges for industrial equipment anomaly detection, such as small-samples problems and significant heterogeneity in the environment. However, this method is very sensitive not easy to manually adjust an appropriate parameter excellent results. To address challenge, paper proposes adaptive personalized federated framework. Specifically, we first design hyperparameters online learning rate adaptation dynamically. Then, a training mechanism with based on...

Point cloud, as a 3D representation, is widely used in autonomous driving, virtual reality (VR), and augmented (AR). However, traditional communication systems think that the point cloud's semantic information irrelevant to communication, which hinders efficient transmission of clouds era artificial intelligence (AI). This paper proposes cloud based system (PCSC), uses AI-based encoding techniques extract joint source-channel coding (JSCC) technology overcome distortion caused by noise...

Federated learning allows for collaborative training of distributed deep networks, tackling the issue data silos while ensuring privacy. However, in setting industrial equipment, collected is dynamic, non-independent, and identically distributed(i.e., Non-IID), which negatively impacts performance federated learning. While personalized solution can address Non-IID problem, it falls short effectively handling dynamic environment poses a significant challenge to model stability. Model...

In this paper, we propose a new wireless video communication scheme to achieve high-efficiency transmission over noisy channels. It exploits the idea of model division multiple access (MDMA) and extracts common semantic features across frames. Besides, deep joint source-channel coding (JSCC) is applied overcome distortion caused by The proposed framework collected under name (MDVSC). particular, temporal relative frames are first transformed into latent space for computing complexity...

Semantic communication, as a revolutionary communication architecture, is considered promising novel paradigm. Unlike traditional symbol-based error-free systems, semantic-based visual systems extract, compress, transmit, and reconstruct images at the semantic level. However, widely used image similarity evaluation metrics, whether pixel-based MSE or PSNR structure-based MS-SSIM, struggle to accurately measure loss of semantic-level information source during system transmission. This...

Air conditioning contributes a high percentage of energy consumption over the world. The efficient prediction can help to reduce consumption. Traditionally, multidimensional air data could only be processed sequentially for each dimension, thus resulting in inefficient feature extraction. Furthermore, due reasons such as implicit correlations between hyperparameters, automatic hyperparameter optimization (HPO) approaches not easily achieved. In this paper, we propose an auto-optimization...

The end-to-end image communication system has been widely studied in the academic community. escalating demands on systems terms of data volume, environmental complexity, and task precision require enhanced efficiency, anti-noise ability semantic fidelity. Therefore, we proposed a novel paradigm based Semantic Feature Decomposition (SeFD) for integration large-scale visual generation models to achieve high-performance, highly interpretable controllable communication. According this paradigm,...

In this article, we propose a comprehensive approach to responsible artificial intelligence (AI)-based software architecture for the digitalization of industry drawings, serving as engineering reference AI in other domains.

This paper introduces a novel method for transmitting video data over noisy wireless channels with high efficiency and controllability. The derivates from model division multiple access (MDMA) to extract common semantic features frames. It also uses deep joint source-channel coding (JSCC) as the main framework establish communication links deal channel noise. An entropy-based variable length scheme is developed adjust amount accurately explicitly. We name our (MDVSC). steps of approach are...

Semantic communication can improve efficiency and quality. However, when the distance is long, it will cause severe signal attenuation. This paper discusses a cooperative relay network based on semantic coding (CR-SNC) to reliability problem in long-distance image transmission. By introducing node, we cooperatively encode information of multiple users at relay. Without reducing transmission efficiency, improved. The solution efficiently utilizes characteristic by using techniques jointly...

Nowadays, the need for high-quality image reconstruction and restoration is more urgent. However, most transmission systems may suffer from quality degradation or interruption in face of interference such as channel noise link fading. To solve this problem, a relay communication network semantic based on shared feature extraction hyperprior entropy compression (HEC) proposed, where technology Pearson correlation proposed to eliminate partial extracted latent feature. In addition, HEC used...

Semantic communication, as a novel communication paradigm, has attracted the interest of many scholars, with multi-user, multi-input multi-output (MIMO) scenarios being one critical contexts. This paper presents semantic importance-aware based system (SIA-SC) over MIMO Rayleigh fading channels. Combining symbols' inequality and equivalent subchannels channels on Singular Value Decomposition (SVD) maximizes end-to-end performance through new layer mapping method. For multi-user scenarios,...