Goto

Collaborating Authors

Google Brain researchers teach AI to make its own encryption

#artificialintelligence

Researchers at Google Brain, Google's deep learning project, have worked out a way to teach artificial intelligence neural networks how to create their own encryption formulas. In a research paper published by Martín Abadi and David Andersen, "Learning to Protect Communications with Adversarial Neural Cryptography," the researchers put three AIs together: two that attempt to communicate secretly (Alice and Bob) and a third that tries to spy on that communication (Eve). According to the paper, Alice's job is to construct messages using some sort of secret algorithm and Bob's duty is to discover how to decrypt those messages. On the other side of the divide, Eve is listening in to the communication between Alice and Bob and attempts each time to read the message sent. The objective of the entire process is to have Alice and Bob come up with a communication scheme that Eve cannot easily break, all without teaching Alice, Bob or Eve any particular encryption scheme.


Google's neural networks invent their own encryption

#artificialintelligence

A team from Google Brain, Google's deep learning project, has shown that machines can learn how to protect their messages from prying eyes. Researchers Martín Abadi and David Andersen demonstrate that neural networks, or "neural nets" – computing systems that are loosely based on artificial neurons – can work out how to use a simple encryption technique. In their experiment, computers were able to make their own form of encryption using machine learning, without being taught specific cryptographic algorithms. The encryption was very basic, especially compared to our current human-designed systems. Even so, it is still an interesting step for neural nets, which the authors state "are generally not meant to be great at cryptography".


A New Kind of AI: Google's Deep Learning Neural Nets Have Learned Encryption

#artificialintelligence

Alice and Bob can keep secrets -- well, at least from Eve. These three are the neural networks (or neural nets) that a team from Google Brain, Google's research division for machine deep learning, developed to see just how well artificial intelligence (AI) can keep secrets. It turns out, they can do it pretty well. In a study published on arXiv, researchers Martín Abadi and David Andersen feature how neural nets can develop their own simple encryption techniques in order to keep messages from eavesdroppers, even without being given special cryptographic algorithms. In theory, neural nets "are generally not meant to be great at cryptography," the researchers said.


A New Kind of AI: Google's Deep Learning Neural Nets Have Learned Encryption

#artificialintelligence

These three are the neural networks (or neural nets) that a team from Google Brain, Google's research division for machine deep learning, developed to see just how well artificial intelligence (AI) can keep secrets. It turns out, they can do it pretty well. In a study published on arXiv, researchers Martín Abadi and David Andersen feature how neural nets can develop their own simple encryption techniques in order to keep messages from eavesdroppers, even without being given special cryptographic algorithms. In theory, neural nets "are generally not meant to be great at cryptography," the researchers said. The experiment involved these three, affectionately named, neural nets.


Google's neural networks invent their own encryption

New Scientist

A team from Google Brain, Google's deep learning project, has shown that machines can learn how to protect their messages from prying eyes. Researchers Martín Abadi and David Andersen demonstrate that neural networks, or "neural nets" – computing systems that are loosely based on artificial neurons – can work out how to use a simple encryption technique. In their experiment, computers were able to make their own form of encryption using machine learning, without being taught specific cryptographic algorithms. The encryption was very basic, especially compared to our current human-designed systems. Even so, it is still an interesting step for neural nets, which the authors state "are generally not meant to be great at cryptography".