Putting neural networks under the microscope

In a study that sheds light on how these systems manage to translate text from one language to another, the researchers developed a method that pinpoints individual nodes, or “neurons,” in the networks that capture specific linguistic features. In machine translation, a network crunches language data annotated by humans, and presumably “learns” linguistic features, such as word morphology, sentence structure, and word meaning. In their paper, the researchers used the method to show that, by ablating high ranking neurons in a network, its performance in classifying correlated linguistic features dipped significantly.

Source: news.mit.edu