Federated Hierarchical Tensor Networks: a Collaborative Learning Quantum AI-Driven Framework for Healthcare

Amandeep Singh Bhatia,David E. Bernal Neira

2024-07-04

Abstract:Healthcare industries frequently handle sensitive and proprietary data, and due to strict privacy regulations, they are often reluctant to share data directly. In today's context, Federated Learning (FL) stands out as a crucial remedy, facilitating the rapid advancement of distributed machine learning while effectively managing critical concerns regarding data privacy and governance. The fusion of federated learning and quantum computing represents a groundbreaking interdisciplinary approach with immense potential to revolutionize various industries, from healthcare to finance. In this work, we proposed a federated learning framework based on quantum tensor networks, which leverages the principles of many-body quantum physics. Currently, there are no known classical tensor networks implemented in federated settings. Furthermore, we investigated the effectiveness and feasibility of the proposed framework by conducting a differential privacy analysis to ensure the security of sensitive data across healthcare institutions. Experiments on popular medical image datasets show that the federated quantum tensor network model achieved a mean receiver-operator characteristic area under the curve (ROC-AUC) between 0.91-0.98. Experimental results demonstrate that the quantum federated global model, consisting of highly entangled tensor network structures, showed better generalization and robustness and achieved higher testing accuracy, surpassing the performance of locally trained clients under unbalanced data distributions among healthcare institutions.

Quantum Physics,Artificial Intelligence,Machine Learning

What problem does this paper attempt to address?

The main problems that this paper attempts to solve are the difficulties of data privacy and sharing in the medical industry. Specifically:

1. **Data Privacy and Regulatory Restrictions**: Data in the medical industry usually contains sensitive and personally identifiable information, and is thus protected by strict privacy regulations. This leads to concerns among medical institutions when directly sharing data, fearing possible infringement of patients' privacy and incurring legal and ethical consequences.

2. **Data Heterogeneity**: There are significant differences in the format, standards, and structure of medical data, which makes it difficult for centralized machine learning to effectively utilize data sources from different institutions.

3. **Data Deficiency and Uneven Distribution**: Small - scale medical institutions may lack sufficient data to train accurate prediction models, especially when dealing with rare diseases. Meanwhile, data is often unevenly distributed among different institutions, which further increases the difficulty of model training.

To solve these problems, the paper proposes a federated learning framework based on quantum tensor networks (FedQTN), which combines quantum computing and collaborative learning techniques, aiming at:

- **Protecting Data Privacy**: Ensure the security and privacy of patients' data by processing data locally and only sharing model updates.

- **Improving Model Performance**: Utilize the efficient computing power of quantum tensor networks to enhance the generalization ability and robustness of the model, especially when dealing with unbalanced data distributions.

- **Promoting Cross - institutional Cooperation**: Enable multiple medical institutions to collaboratively train models without directly sharing raw data, thereby improving the overall performance of the model.

The paper verifies the effectiveness of this framework through experiments on multiple medical image datasets, demonstrating its excellent performance in tasks such as chest X - ray images, brain MRI scans, and kidney CT scans.