In recent years, powered by state-of-the-art achievements in a broad range of areas, machine learning has received considerable attention from the healthcare sector. Despite their ability to provide solutions within personalized medicine, strict regulations on the confidentiality of patient health information have in many cases hindered the adoption of deep learning-based solutions in clinical workflows. To allow for the processing of sensitive health information without disclosing the underlying data, we propose a solution based on fully homomorphic encryption (FHE). The considered encryption scheme, MORE (Matrix Operation for Randomization or Encryption), enables the computations within a neural network model to be directly performed on floating point data with a relatively small computational overhead. We consider the well-known MNIST digit recognition problem to evaluate the feasibility of the proposed method and show that performance does not decrease when deep learning is applied on MORE homomorphic data. To further evaluate the suitability of the method for healthcare applications, we first train a model on encrypted data to estimate the outputs of a whole-body circulation (WBC) hemodynamic model and then provide a solution for classifying encrypted X-ray coronary angiography medical images. The findings highlight the potential of the proposed privacy-preserving deep learning methods to outperform existing approaches by providing, within a reasonable amount of time, results equivalent to those achieved by unencrypted models. Lastly, we discuss the security implications of the encryption scheme and show that while the considered cryptosystem promotes efficiency and utility at a lower security level, it is still applicable in certain practical use cases.

Data privacy is a major concern when accessing and processing sensitive medical data. A promising approach among privacy-preserving techniques is homomorphic encryption (HE), which allows for computations to be performed on encrypted data. Currently, HE still faces practical limitations related to high computational complexity, noise accumulation, and sole applicability the at bit or small integer values level. We propose herein an encoding method that enables typical HE schemes to operate on real-valued numbers of arbitrary precision and size. The approach is evaluated on two real-world scenarios relying on EEG signals: seizure detection and prediction of predisposition to alcoholism. A supervised machine learning-based approach is formulated, and training is performed using a direct (non-iterative) fitting method that requires a fixed and deterministic number of steps. Experiments on synthetic data of varying size and complexity are performed to determine the impact on runtime and error accumulation. The computational time for training the models increases but remains manageable, while the inference time remains in the order of milliseconds. The prediction performance of the models operating on encoded and encrypted data is comparable to that of standard models operating on plaintext data.

The industrial environment has gone through the fourth revolution, also called “Industry 4.0”, where the main aspect is digitalization. Each device employed in an industrial process is connected to a network called the industrial Internet of things (IIOT). With IIOT manufacturers being capable of tracking every device, it has become easier to prevent or quickly solve failures. Specifically, the large amount of available data has allowed the use of artificial intelligence (AI) algorithms to improve industrial applications in many ways (e.g., failure detection, process optimization, and abnormality detection). Although data are abundant, their access has raised problems due to privacy concerns of manufacturers. Censoring sensitive information is not a desired approach because it negatively impacts the AI performance. To increase trust, there is also the need to understand how AI algorithms make choices, i.e., to no longer regard them as black boxes. This paper focuses on recent advancements related to the challenges mentioned above, discusses the industrial impact of proposed solutions, and identifies challenges for future research. It also presents examples related to privacy-preserving and explainable AI solutions, and comments on the interaction between the identified challenges in the conclusions.

Deep learning (DL)-based algorithms have demonstrated remarkable results in potentially improving the performance and the efficiency of healthcare applications. Since the data typically needs to leave the healthcare facility for performing model training and inference, e.g., in a cloud based solution, privacy concerns have been raised. As a result, the demand for privacy-preserving techniques that enable DL model training and inference on secured data has significantly grown. We propose an image obfuscation algorithm that combines a variational autoencoder (VAE) with random non-bijective pixel intensity mapping to protect the content of medical images, which are subsequently employed in the development of DL-based solutions. A binary classifier is trained on secured coronary angiographic frames to evaluate the utility of obfuscated images in the context of model training. Two possible attack configurations are considered to assess the security level against artificial intelligence (AI)-based reconstruction attempts. Similarity metrics are employed to quantify the security against human perception (structural similarity index measure and peak signal-to-noise-ratio). Furthermore, expert readers performed a visual assessment to determine to what extent the reconstructed images are protected against human perception. The proposed algorithm successfully enables DL model training on obfuscated images with no significant computational overhead while ensuring protection against human eye perception and AI-based reconstruction attacks. Regardless of the threat actor’s prior knowledge of the target content, the coronary vessels cannot be entirely recovered through an AI-based attack. Although a drop in accuracy can be observed when the classifier is trained on obfuscated images, the performance is deemed satisfactory in the context of a privacy–accuracy trade-off.

Medical wearable devices monitor health data and, coupled with data analytics, cloud computing, and artificial intelligence (AI), enable early detection of disease. Privacy issues arise when personal health information is sent or processed outside the device. We propose a framework that ensures the privacy and integrity of personal medical data while performing AI-based homomorphically encrypted data analytics in the cloud. The main contributions are: (i) a privacy-preserving cloud-based machine learning framework for wearable devices, (ii) CipherML—a library for fast implementation and deployment of deep learning-based solutions on homomorphically encrypted data, and (iii) a proof-of-concept study for atrial fibrillation (AF) detection from electrocardiograms recorded on a wearable device. In the context of AF detection, two approaches are considered: a multi-layer perceptron (MLP) which receives as input the ECG features computed and encrypted on the wearable device, and an end-to-end deep convolutional neural network (1D-CNN), which receives as input the encrypted raw ECG data. The CNN model achieves a lower mean F1-score than the hand-crafted feature-based model. This illustrates the benefit of hand-crafted features over deep convolutional neural networks, especially in a setting with a small training data. Compared to state-of-the-art results, the two privacy-preserving approaches lead, with reasonable computational overhead, to slightly lower, but still similar results: the small performance drop is caused by limitations related to the use of homomorphically encrypted data instead of plaintext data. The findings highlight the potential of the proposed framework to enhance the functionality of wearables through privacy-preserving AI, by providing, within a reasonable amount of time, results equivalent to those achieved without privacy enhancing mechanisms. While the chosen homomorphic encryption scheme prioritizes performance and utility, certain security shortcomings remain open for future development.