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

Developing Its Own Encryption

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.

The experiment involved these three, affectionately named, neural nets. The idea was for Alice and Bob to pass messages to each other while Eve tries to intercept and decode the exchanges. In order to do this, Alice and Bob were given a pre-determined set of numbers as a key to help encrypt and decrypt the message. Eve had no access to this key.

Credits: Abadi and Andersen/Google Brain

After 15,000 iterations of the scenario, Alice and Bob became adept at developing their own simple encryption technique. Alice managed to convert the original plain-text into a cipher text that Bob, in turn, was able to decrypt. Eve only managed to decipher 8 of the 16-bits in the message, where each bit was either a 1 or 0 — a success rate equivalent to pure luck.

AI is everywhere

Google’s work on deep learning has already made considerable development in AI ability to interpret speech and recognize images (with Cloud Vision), and even one that’s capable of human-level speech. Even more, it’s not just limited to practical tasks. A Google AI has tried exploring the arts by making its own song.

So, while AI has only just started to try out cryptography, it has gradually been in, pretty much, everything else.

Neural nets are capable of all this because they are computer networks modeled after the human brain. This is what’s fascinating with AI aggregate technologies, like deep learning. It keeps getting better, learning on its own, with some even capable of self training.

We are, indeed, at the center of one of the biggest moments in our civilization’s history.

The post A New Kind of AI: Google’s Deep Learning Neural Nets Have Learned Encryption appeared first on Futurism.

Source: <a href="A New Kind of AI: Google’s Deep Learning Neural Nets Have Learned Encryption« >TheOldReader

Laisser un commentaire

Votre adresse de messagerie ne sera pas publiée. Les champs obligatoires sont indiqués avec *