Secure multi-party linear regression at plaintext speed

  • 2019-07-10 17:57:00
  • Jonathan M. Bloom
  • 0

Abstract

We detail distributed algorithms for scalable, secure multiparty linearregression and feature selection at essentially the same speed as plaintextregression. While the core geometric ideas are simple, the recognition of theirbroad utility when combined is novel. Our scheme opens the door to efficientand secure genome-wide association studies across multiple biobanks.

 

Quick Read (beta)

This feature is not avaialbe for this paper.