What's in Your Embedding, And How It Predicts Task Performance
Attempts to find a single technique for general-purpose intrinsic evaluation of word embeddings have so far not been successful. We present a new approach based on scaled-up qualitative analysis of word vector neighborhoods that quantifies interpretable characteristics of a given model (e.g. its preference for synonyms or shared morphological forms as nearest neighbors). We analyze 21 such factors and show how they correlate with performance on 14 extrinsic and intrinsic task datasets (and also explain the lack of correlation between some of them). Our approach enables multi-faceted evaluation, parameter search, and generally {--} a more principled, hypothesis-driven approach to development of distributional semantic representations.
PDF Abstract COLING 2018 PDF COLING 2018 Abstract
SNLI
