Journal of Ubiquitous Computing and Communication Technologies
Abstract
Early prediction of cardiovascular diseases has always been a significant challenge due to data privacy risks, non-interpretable outcomes, and demographic biases associated with current machine learning approaches. In this paper, CARE-XAI (Comprehensive AI Risk Evaluation with Explainable Artificial Intelligence) is presented as a novel framework that combines federated learning, fairness optimization methods, and explainable artificial intelligence to predict cardiovascular diseases based on multiple clinical modalities. CARE-XAI uses the UCI Heart Disease database as the fundamental clinical modality, along with simulated behavioral and wearable datasets to improve predictive accuracy. Federated learning using the FedAvg algorithm ensures privacy-preserving training in a decentralized manner without any leakage of sensitive information about patients. Fairness optimization with the help of AIF360 is used to counteract demographic biases, while model transparency is guaranteed through SHAP explanations. Experimental results show that CARE-XAI provides an accuracy score of 83.1% with 82.5% precision, 84.3% recall, and 0.881 AUC values, while also performing comparably well with centralized frameworks.
References
- Krittanawong, Chayakrit, HongJu Zhang, Zhen Wang, Mehmet Aydar, and Takeshi Kitai. "Artificial intelligence in precision cardiovascular medicine." Journal of the American College of Cardiology 69, no. 21 (2017): 2657-2664.
- Dey, Damini, Piotr J. Slomka, Paul Leeson, Dorin Comaniciu, Sirish Shrestha, Partho P. Sengupta, and Thomas H. Marwick. "Artificial intelligence in cardiovascular imaging: JACC state-of-the-art review." Journal of the American College of Cardiology 73, no. 11 (2019): 1317-1335.
- Rajkomar, Alvin, Jeffrey Dean, and Isaac Kohane. "Machine learning in medicine." New England Journal of Medicine 380, no. 14 (2019): 1347-1358.
- Esteva, Andre, Katherine Chou, Serena Yeung, Nikhil Naik, Ali Madani, Ali Mottaghi, Yun Liu, Eric Topol, Jeff Dean, and Richard Socher. "Deep learning-enabled medical computer vision." NPJ digital medicine 4, no. 1 (2021), p. 5.
- Johnson, Kipp W., Jessica Torres Soto, Benjamin S. Glicksberg, Khader Shameer, Riccardo Miotto, Mohsin Ali, Euan Ashley, and Joel T. Dudley. "Artificial intelligence in cardiology." Journal of the American College of Cardiology 71, no. 23 (2018): 2668-2679.
- Alaa, Ahmed M., Thomas Bolton, Emanuele Di Angelantonio, James HF Rudd, and Mihaela Van der Schaar. "Cardiovascular disease risk prediction using automated machine learning: A prospective study of 423,604 UK Biobank participants." PloS one 14, no. 5 (2019): e0213653.
- Xu, Jie, Benjamin S. Glicksberg, Chang Su, Peter Walker, Jiang Bian, and Fei Wang. "Federated learning for healthcare informatics." Journal of healthcare informatics research 5, no. 1 (2021): 1-19.
- Warnat-Herresthal, Stefanie, Hartmut Schultze, Krishnaprasad Lingadahalli Shastry, Sathyanarayanan Manamohan, Saikat Mukherjee, Vishesh Garg, Ravi Sarveswara "Swarm learning for decentralized and confidential clinical machine learning." Nature 594, no. 7862 (2021): 265-270.
- Li, Tian, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, and Virginia Smith. "Federated optimization in heterogeneous networks." Proceedings of Machine learning and systems 2 (2020): 429-450.
- Kairouz, Peter, and H. Brendan McMahan. "Advances and open problems in federated learning." Foundations and trends in machine learning 14, no. 1-2 (2021): 1-210.
- Janosi, A., Steinbrunn, W., Pfisterer, M., & Detrano, R. Heart Disease [Dataset]. UCI Machine Learning Repository - https://archive.ics.uci.edu/dataset/45/heart+disease