Learning to Protect Communications with Adversarial Neural Cryptography

TL;DR

This paper explores whether neural networks can learn to encrypt and decrypt information to protect communication confidentiality in a multiagent system through adversarial training.

Contribution

It introduces a method for neural networks to learn encryption and decryption end-to-end without predefined cryptographic algorithms, focusing on confidentiality against adversaries.

Findings

Neural networks can learn encryption and decryption functions.

Networks can selectively apply cryptographic operations to ensure confidentiality.

Adversarial training enables neural networks to develop secure communication protocols.

Abstract

We ask whether neural networks can learn to use secret keys to protect information from other neural networks. Specifically, we focus on ensuring confidentiality properties in a multiagent system, and we specify those properties in terms of an adversary. Thus, a system may consist of neural networks named Alice and Bob, and we aim to limit what a third neural network named Eve learns from eavesdropping on the communication between Alice and Bob. We do not prescribe specific cryptographic algorithms to these neural networks; instead, we train end-to-end, adversarially. We demonstrate that the neural networks can learn how to perform forms of encryption and decryption, and also how to apply these operations selectively in order to meet confidentiality goals.