Privacy-preserving semi-parallel logistic regression training with fully homomorphic encryption

摘要

The advent of next generation sequencing and the progressive reduction of costs in sequencing processes result in an increasing amount of available genomic data, which is essential for better modeling the relation between genotypic traits, predisposition to diseases, response to treatments, effect of drugs, and, in general, for achieving more accurate models that enable personalized and precision medicine. While machine learning computations on large scale genomic data present obvious outsourcing needs and can benefit from Cloud services, the high sensitivity of genomic data and the impossibility of properly anonymizing it [ , ] call for effective protection methods that enable accurate and efficient computation without leaking information about the individual genomic sequences to untrusted cloud service providers. In order to become feasible and usable for the purpose of personalized medicine, these protection mechanisms must optimize the trade-off between the accuracy of the results, the efficiency of the computation, and the security level.