Scientific Conferences of Ukraine, 6th International Conference High Performance Computing HPC-UA 2020

Font Size: 
Blockchain and Smart Contracts for Provenance of Deep Learning Content in Healthcare
Yuri Gordienko, Oleg Alienin, Oleksandr Rokovyi, Volodymyr Valko, Sergii Stirenko, Gilles Fedak, Oleg Lodygensky

Last modified: 2021-10-18

Abstract


Due to advances of artificial intelligence (AI) and deep learning (DL) techniques, the opportunities for the reliable medical classification and prediction of some diseases become possible in recent years. Some predictions made by DL neural network trained on the huge medical datasets (MD) sometimes overcome the experts in the field and DL-models can be considered as useful AI-screening tools and good assistant for the real doctors. In this context, the proof of authenticity (PoA) of such deep learning content (DLC) (like datasets, models, etc.) is very important to realize the origin and evolution of DLC. At the moment there are no convenient solutions that can provide history tracking and provenance of DLC. In this paper, we provide a general framework using Ethereum smart contracts to track back the provenance and evolution of DLC to its original source even if the DLC was edited (e.g. DL models were retrainedor/and datasets were updated) by anonymous authors.The main principle behind the solution is that if the DLC can be credibly traced to a trusted or reputable source, the DLC can then be real and authentic. The solution is proposed in the healthcare context and for medical DLC, but it can be applied to any other form of DLC.

Full Text: PDF