Group key exchange using neural cryptography with binary trees

Key exchange is one of the major concerns in cryptology. Mutual learning between two neural networks has been used to address this issue. However, there is still a need to extend the concept so that a group of parties can share a common key. This work suggests an algorithm that employs and ex- tends the mutual learning process to accommodate the much needed group secure communication. Here, a recursive algorithm that depends on ordering the parties in a binary tree structure is introduced. By applying this algorithm, a group of N parties is able to share a common key in a time proportional to sup (log₂(N)). The proposed algorithm implements two binary tree approaches to achieve this objective, one uses a binary tree with swapping, BTWS, while the other employs a binary tree with election, BTWE.