Disclosure control of machine learning models from trusted research environments (TRE): New challenges and opportunities. Issue 4 (April 2023)

• Record Type: Journal Article
• Authors: Mansouri-Benssassi, Esma; Rogers, Simon; Reel, Smarti; Malone, Maeve; Smith, Jim; Ritchie, Felix; Jefferson, Emily
• Is Part Of: Heliyon. Volume 9:Issue 4(2023)
• Publication Date: 2023-04
• Subjects: Trusted research environment -- Safe haven -- AI -- Machine learning -- Data privacy -- Disclosure control; AI Artificial Intelligence -- BBN Bayesian Belief Network -- BERT Bidirectional Encoder Representations from Transformers -- CHAID Chi-squared Automatic Interaction Detection -- CNN Convolution Neural Network -- CT Computerised Tomography -- DBN Deep Belief Network -- EHR Electronic Health Records -- GDPR General Data Protection Regulation -- KNN K-nearest neighbor -- LSTM Long Short-Term Memory networks -- MDS Multi-dimensional Scaling -- MEI Membership Inference -- ML Machine Learning -- MOI MOdel Inversion attacks -- PCA Principal Component Analysis -- PHTs Personalised Health Trains -- RL Reinforcement Learning -- RNN Recurrent Neural Network -- SOM Self-Organising Map -- SSL Self-supervised learning -- SVM Support Vector Machines -- TRE Trusted Research Environment -- TSNE T-Stochastic Neighbour Embedding; Research -- Periodicals; Medical sciences -- Periodicals; Natural history -- Periodicals; Social sciences -- Periodicals; Earth sciences -- Periodicals; Physical sciences -- Periodicals; 507.2
• DOI: 10.1016/j.heliyon.2023.e15143
• Languages: English
• ISSNs: 2405-8440

Abstract

Abstract: Introduction: Artificial intelligence (AI) applications in healthcare and medicine have increased in recent years. To enable access to personal data, Trusted Research Environments (TREs) (otherwise known as Safe Havens) provide safe and secure environments in which researchers can access sensitive personal data and develop AI (in particular machine learning (ML)) models. However, currently few TREs support the training of ML models in part due to a gap in the practical decision-making guidance for TREs in handling model disclosure. Specifically, the training of ML models creates a need to disclose new types of outputs from TREs. Although TREs have clear policies for the disclosure of statistical outputs, the extent to which trained models can leak personal training data once released is not well understood. Background: We review, for a general audience, different types of ML models and their applicability within healthcare. We explain the outputs from training a ML model and how trained ML models can be vulnerable to external attacks to discover personal data encoded within the model. Risks: We present the challenges for disclosure control of trained ML models in the context of training and exporting models from TREs. We provide insights and analyse methods that could be introduced within TREs to mitigate the risk of privacy breaches when disclosing trained models. Discussion: Although specific guidelines and policies exist for statistical disclosure controls in TREs, they do not satisfactorily address these new types of output requests; i.e., trained ML models. There is significant potential for new interdisciplinary research opportunities in developing and adapting policies and tools for safely disclosing ML outputs from TREs.