ACS: Accuracy-based client selection mechanism for federated industrial IoT

Putra, Made Adi Paramartha; Putri, Adinda Riztia; Zainudin, Ahmad; Kim, Dong-Seong; Lee, Jae-Min

doi:10.1016/j.iot.2022.100657

Cited by 28 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, the evaluation of the proposed ADCT is conducted. The performance is compared with state-of-the-art methods such as FedAvg [3], FedMedian [21], FedACS [9], FedYogi [22], and FedOpt [22]. Firstly, the final model accuracy at the end of the federation process is investigated.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Federated Learning Performance through Adaptive Client Optimization with Hyperparameter Tuning

Putra

2024

J. Appl. Data Sci.

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

“…FL optimization can be divided into several aspect, such as client selection [9], local optimization, as well as the security enhancement. The local training process on each selected client is a crucial factor in ensuring the performance of FL.…”

Section: Related Workmentioning

confidence: 99%

Enhancing Federated Learning Performance through Adaptive Client Optimization with Hyperparameter Tuning

Putra

2024

J. Appl. Data Sci.

View full text Add to dashboard Cite

“…From other hand, client selection for training process in federated edge computing has its own limitations like difficulty to address all clients, prolong convergence time, and lower in update model performance in each iteration of training process [40]. In such cases, the incentive scheme is crucial to encouraging and incentivizing FECparticipating nodes to actively and trustworthy provide their data and processing resources to the FEC environment for training purposes.…”

Section: Incentive Scheme and Challengesmentioning

confidence: 99%

Federated Learning Incentivize with Privacy-Preserving for IoT in Edge Computing in the Context of B5G

Jalali,

Chen

2024

Preprint

View full text Add to dashboard Cite

Federated learning and edge computing have resulted in the broad adoption of internet of Things (IoT) due to their fast reaction times and low connection costs. In general, edge computing requires users to send raw data to a central server for further processing. However, this data frequently contains sensitive information that individuals may not want to disclose. As a result, transferring user data with sensitive information increases the risk of data leakage when various unauthorized devices access it. Integrating federated learning with edge computing improves privacy by creating a consistent deep learning model across devices, eliminating the need for real data transferring but the complexity and heterogeneity of the IoT environment present challenges like privacy, low communication, incentives, and seamless data aggregation. In this work, we concentrate on improving privacy with communication stability and designing incentive mechanisms to motivate more clients to participate in the model training process to enhance the performance and accuracy of data aggregation in a highly trusted environment. As a result, the novelty point is the development of a framework that combined a blockchain with federated edge computing in the context of beyond 5G to address the aforementioned challenges and provide excellent communicative and trusted environment. The study's rigorous evaluation showed that the integration of blockchain and B5G technology significantly improved the overall process of federated edge computing including increased accuracy, prevented loss, and motivated more clients to participate in the training process.

show abstract

“…One of the primary concerns is data privacy, as sensitive data from individual devices could be potentially exposed during the model aggregation process [42]. Further, device authentication becomes crucial to prevent malicious devices from participating in the learning process and injecting false updates, which could adversely affect the global model's performance [43]. Secure communication between the edge devices and the server is another critical issue.…”

Section: Focuses On Edge Computing and Blockchain Integrationmentioning

confidence: 99%

Privacy-Preserving Big Data Security for IoT With Federated Learning and Cryptography

Awan,

Din,

Almogren

et al. 2023

IEEE Access

View full text Add to dashboard Cite

In the ever-expanding Internet of Things (IoT) domain, the production of data has reached an unparalleled scale. This massive data is processed to glean invaluable insights, accelerating a myriad of decision-making processes. Nevertheless, the privacy and security of such information present formidable challenges. This study proposes an innovative methodology for resolving these challenges, by augmenting the privacy and efficacy of big data analytics through federated learning in the IoT ecosystem. The proffered approach amalgamates a hierarchical structure, a scalable learning rate, and a rudimentary cryptographic mechanism to foster learning while ensuring robust privacy and security. Additionally, we introduce a novel communication protocol -SEPP-IoT, designed to facilitate efficient, secure, and confidential interactions between IoT devices and a central server. In our pursuit of optimizing communication overhead, we propose an adaptive data compression algorithm, aimed at curbing the volume of data transferred between IoT devices and the central server. To fortify resilience and fault tolerance, our approach incorporates multiple mechanisms such as data replication, error correction codes, and proactive fault detection and recovery. Trust management, a salient feature of our framework, bolsters the security and integrity of federated learning. We recommend a unique technique that gauges the dependability of IoT nodes using four trust parameters. We employ the FedSim simulator to evaluate our method's effectiveness. The results indicate a notable enhancement in privacy and efficiency of big data analytics within the IoT.

show abstract

ACS: Accuracy-based client selection mechanism for federated industrial IoT

Cited by 28 publications

References 15 publications

Enhancing Federated Learning Performance through Adaptive Client Optimization with Hyperparameter Tuning

Enhancing Federated Learning Performance through Adaptive Client Optimization with Hyperparameter Tuning

Federated Learning Incentivize with Privacy-Preserving for IoT in Edge Computing in the Context of B5G

Privacy-Preserving Big Data Security for IoT With Federated Learning and Cryptography

Contact Info

Product

Resources

About