[22] Graham Neubig - Unsupervised Learning of Lexical Information
The Thesis Review
00:00
The Uses of Formalisms in Machine Learning
The formalism itself is actually relatively easily interpretable and manipulable by a human actor who wants to interface with a system. So for example, linguists who are looking to build an analyzer of word forms or something like this, the very first thing they do will not be generate lots of training data. But I think they're particularly useful if you want to essentially interface with some sort of human curator or something like that. And I've actually used them in that way in some of my work.
Transcript
Play full episode
