2020/03/23 by Alireza Nooraiepour, Nooraiepour, Alireza, Sina Rezaei Aghdam +1
Biochemistry, Genetics and Molecular Biology · Computer Science · Engineering · #Cryptography and Security (cs.CR) #DNA and Biological Computing #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Wireless Communication Security Techniques #Wireless Signal Modulation Classification
paper · pdf · doi:10.48550/arxiv.2003.10577
openalex publication_date 2020/03/23 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Finite-length codes are learned for the Gaussian wiretap channel in an\nend-to-end manner assuming that the communication parties are equipped with\ndeep neural networks (DNNs), and communicate through binary phase-shift keying\n(BPSK) modulation scheme. The goal is to find codes via DNNs which allow a pair\nof transmitter and receiver to communicate reliably and securely in the\npresence of an adversary aiming at decoding the secret messages. Following the\ninformation-theoretic secrecy principles, the security is evaluated in terms of\nmutual information utilizing a deep learning tool called MINE (mutual\ninformation neural estimation). System performance is evaluated for different\nDNN architectures, designed based on the existing secure coding schemes, at the\ntransmitter. Numerical results demonstrate that the legitimate parties can\nindeed establish a secure transmission in this setting as the learned codes\nachieve points on almost the boundary of the equivocation region.\n