3-Party Adversarial Cryptography

The domain of Artificial Intelligence (AI) has seen an outstanding growth during the last two decades. It has proven its efficiency in handling complex domains including speech recognition, image recognition and many more. One interesting and evolving branch that was put forward years ago but have seen a good growth only during the past few years is encryption using AI. After Google announced that it has succeeded teaching neural networks encryption in the presence of Eavesdroppers, research in this particular area has seen a rapid spread of interest among different researchers all over the world to develop new Neural Networks capable of operating different cryptographic tasks. In this paper, we take initial steps to achieve secure communication among more than two parties using neural network based encryption. We forward the idea of two party symmetric encryption scheme of Google to a multi party Encryption scheme. In this paper we will focus on a 3-Party case.