Sandi Rahmadika, Winda Agustiarmi, Delsina Faiza, Putra Jaya, Thamrin, Ryan Fikri
Collaborative learning techniques facilitate the collective participation of multiple individuals in enhancing artificial intelligence models by leveraging their respective private datasets. The training is conducted by the users in a local setting, with gadgets facilitating the frequent exchange of gradient values. In contrast to traditional training methods, collaborative procedures do not involve publicly disclosing training data. Irrespective of the positives associated with privacy concerns, customers frequently exhibit reduced motivation to enhance the model owing to insufficient procedural incentives. In summary, the available resources are not utilized to their fullest potential. In order to address the problem at hand, we have developed a collaborative learning model that incorporates a secure, equitable, and unalterable reward mechanism through the use of blockchain technology. Incentives are allocated in a manner that is commensurate with the individual contributions made by users. The incentive schemes are implemented on the Ethereum blockchain. Additionally, we assess the efficacy of collaborative learning in an alternative context. The experimental results show that the design objectives have been achieved. © 2023 IEEE.
Universitas Negeri Padang, Faculty of Engineering, Department of Electronic Engineering, Sumatera Barat, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock