Learning to Protect Communications with Adversarial Neural Cryptography (nonfiction)

Neural network cryptography diagram.

Learning to Protect Communications with Adversarial Neural Cryptography is a research paper by Martín Abadi, David G. Andersen of Google Brain.

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.

It was published in October 2016.

Nonfiction cross-reference

• Artificial intelligence (nonfiction)
• Cryptography (nonfiction)

External links:

• Learning to Protect Communications with Adversarial Neural Cryptography @ Cornell University Library
• Google's AI created its own form of encryption
• Google's AI Created Its Own Form of Encryption @ Slashdot