Oct

21st

Mon
21st

How Google Converted Language Translation Into a Problem of Vector Space Mathematics | MIT Technology Review

This reminds me of Sapir-Whorf, combined with the idea that concepts are somehow “eternal” and reflected by the structural relations one concept has with another, with the result that conceptual structures embedded in vocabularies across languages can often map on to each other.

The new trick is to represent an entire language using the relationship between its words. The set of all the relationships, the so-called “language space”, can be thought of as a set of vectors that each point from one word to another. And in recent years, linguists have discovered that it is possible to handle these vectors mathematically. For example, the operation ‘king’ – ‘man’ + ‘woman’ results in a vector that is similar to ‘queen’.