Amazing discoveries

Computers that create their own secret language. The end of black and white photographs. And an algorithm that knows what will come next in your photo. In this article three new examples of skills that computers have and humans not so much.

1. Esperanto was once invented to help people from different cultures to talk to each other using an easy to learn, politically neutral, language. Recently researchers at Google witnessed the convergence of a similar concept when machines were -together- trying to talk in a new language.

lingua

TechCrunch couldn’t resist to hype this story a little bit:

GNMT’s creators were curious about something. If you teach the translation system to translate English to Korean and vice versa, and also English to Japanese and vice versa… could it translate Korean to Japanese, without resorting to English as a bridge between them?

As it turns out — yes! It produces “reasonable” translations between two languages that it has not explicitly linked in any way. Remember, no English allowed. Read the full article here, or simply download the 1611-04558v1 here.

2. Colorize
Somethings will never change. Like old black-and-white photographs, right? Wrong. Meet the Colorize algorithm. It uses deep learning to automatically colorize black and white photos—as a prime example.

Colorize was built at the UC Berkeley Vision Lab, as part of a graduate student’s academic paper, and the Algorithmia team urged him to upload the algorithm to their marketplace.

3. Your next move
Imagine if your favorite picture could automatically be converted into a short video and labeled. Sound like a fantasy? Maybe not for long.

Using a deep learning algorithm, MIT’s Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba recently generated one second of predictive video based on a single still frame.

Called Scene Dynamics, the software has been taught with roughly two million unlabeled videos. After being fed a new image, the system runs two competing neural networks. The first generates the predictive video while the second discerns if the videos are real or fake. Beyond predicting an impressive number of frames based on assumed motion, the algorithm also classifies the specific action occurring. While clearly not perfect, the results are impressive already.

Category: algorithms, Research

Discover more: > > >