Blockchain-Powered Tensor Meta-Learning-Driven Intelligent Healthcare System With IoT Assistance

Abstract

The rapid development and gradual integration of artificial intelligence and the Internet of Things have brought unprecedented opportunities for radically changing healthcare and treatments. However, the burgeoning in intelligent healthcare systems is severely bounded by data privacy and the security of AI models. Meanwhile, the limited local data forces conventional AI models to face the predicament in achieving personalized healthcare. Hence, we propose a blockchain-powered tensor meta-learning-driven intelligent healthcare system with IoT assistance. IoT devices as light nodes upload the local shareable data to the edge server(full node) for model training and perform the local private data by non-tampered model downloaded via smart contract. The system can not only use blockchain technology to ensure the strong consistency of the healthcare model but also protect private data from being leaked. Especially, we develop a tensor meta-learning model named tensor-prototype graph network to achieve efficient modeling of heterogeneous healthcare data. Building on the tensors and graph network, the model is conducive to capturing the data distribution when there are few labeled data. To evaluate our proposed approach, we have conducted experiments on three classic databases. The results demonstrate that our approach is capable of effectively promoting the performance of intelligent healthcare.