In the life sciences field, enormous amounts of essential data exist, often containing sensitive information. This data could encompass anything from patient clinical records and genomics to clinical trial recruitment and AI model/analytics. These types of information are very attractive to cyber threats, and the consequences of any breaches in this data are enormous, bringing about both monetary losses and regulatory compliance issues. Safety is a top priority when analyzing and computing on these datasets.
Joint development in this industry proves to be extremely difficult due to concerns about data protection like HIPAA and GDPR in tandem with regulatory compliance. Complications stemming from data sharing have made physical data collection the standard practice when in need of a dataset. This creates an industry-wide need for secured solutions or technologies that safely enable sophisticated analytics and model building. These security for life sciences solutions must be able to access and store sensitive data without data centralization to facilitate industry collaboration.
Recently, Duality has partnered with Intel to pursue this issue and advance the possibility of privacy preserving data collaboration using Privacy Enhancing Technologies (PETs) to enable more secure data analysis and AI on encrypted data. The goal was to tackle real world evidence studies to aid and accelerate medical and pharmaceutical innovation. Real world evidence studies have many concerns surrounding data privacy, security, collaboration, and confidentiality, making collaboration on medical data extremely cumbersome.
To alleviate this issue, Duality created a solution using Intel’s 3rd Generation Intel® Xeon® Scalable processors with built-in AI plus encryption accelerators. Using fully homomorphic encryption on Intel platforms, privacy preserving data collaboration has become possible and scalable, allowing enterprises to safely and effectively compute on sensitive medical data.
Leveraging Federated Learning alongside Homomorphic Encryption, Duality has allowed the centralized orchestration of sensitive data while preventing leakage of intermediate federated model training results. This allows linking patient health records but also doing statistical analyses, regression models, and more on this data while working at speeds faster than most cutting-edge technologies. In this solution, data owners remain in control of how and when their data is used during collaboration. Using Duality’s platform, customers can unleash the full value of their data and AI while minimizing risk by making privacy-enhanced data collaboration advantageous to analytic transformations in the life sciences industry.
Security will continue to be among the top concerns of life sciences organizations. With Duality and Intel’s recent project, there is now a solution to secure data collaboration and customers can harness big data using advanced AI techniques, offering valuable data and analytics to the medical field. If taken advantage of, this technology will permit further innovation and discovery in the field.
To read the full brief, click here.
