Humans have been storing, retrieving, manipulating, and communicating information since the Sumerians in Mesopotamia developed writing in 3000 BCE. Since then, we have continuously developed more and more sophisticated means to communicate and push information. Whether unconsciously or consciously, we seem to always need more data, faster than ever. And with every technological breakthrough that comes along, we also have a set of new concepts that reshape our world.

We can think back, for example, to Gutenberg’s printing press. Invented in 1440, it pushed printing costs down and gave birth to revolutionary concepts like catalogs (the first was published in 1495 in Venice by publisher Aldus Manutiu and listed all the books that he was printing), mass media (which enabled revolutionary ideas to transcend borders), magazines, newspapers, and so on. All these concepts emerged from a single “master” technology breakthrough and have had a great impact on every single aspect of individuals’ lives and the global world picture.

A hundred years later, the core idea of data distribution has not changed much. We still browse catalogs to buy our next pair of shoes, we create catalogs to sell our products and services, and we still browse publications looking for information.

In groundbreaking work on morality-proofing AI, researchers at MIT are designing neural networks that will provide explanations for why they reached a conclusion.

Center for Data Science Professor Kyunghyun Cho talks improving multi-way, multilingual translations.

Although machines can outperform humans in almost any skill set today, there is still one process that they have yet to master: translation. Several students learning a second or third language in particular will have undoubtedly encountered some of the more hilarious results produced by Google (mis)Translate.

But a solution was recently proposed by the Center for Data Science’s very own Kyunghyun Cho. Together with Yoshua Bengio and Orhan Firat, their innovative model — which is the first to handle multi-way, multilingual translations — clinched the runners-up position for best paper at the 2016 Annual Conference of the North American Chapter of the Association for Computational Linguistics.