On the Synergy of Homomorphic Blockchain and Deep Neural Network for Enhanced Data Security: A Review
K. Swanthana, S. S. Aravinth · 2024
Data Security is a concern as we are in an age where the data becomes digitized and can be easily compromised more often being surrounded by cybercrime threats. Meanwhile, researchers have been investigating to solve this issue using various techniques and methods like differential privacy, federated learning, homomorphic encryption. The purpose of these techniques is to improve data security and at the same time, provide effective machine learning and data analysis. Utilizing a homomorphic blockchain approach, data can be encrypted and processed with complete security without compromising confidentiality of any kind by using deep neural network. This is possible through combination of following the best practice in AI using deep neural networks which have been very successful at solving difficult problems and following model security principles by leveraging homomorphic encryption as well as blockchain technology. These models will store the data in the cloud securely, fostering decentralized sharing that provides transparency and traceability. By following this framework with data analytics, companies can still leverage the power of their stored information while maintaining privacy and security over sensitive material. It also protects the data from malicious attacks and encrypted computation can keep both model-training and running over a set of secret databases in an almost fully-encrypted fashion. As a result, the risk of data leakage or abuse decreases and an additional security barrier is established.