All-but-the-Top: Simple and Effective Postprocessing for Word Representations
This incremental improvement addresses the need for more effective word embeddings in NLP applications.
The paper tackles the problem of improving off-the-shelf word representations by introducing a simple postprocessing technique that removes the common mean vector and top dominating directions, resulting in consistently better performance across various lexical and sentence-level tasks in multiple languages.
Real-valued word representations have transformed NLP applications; popular examples are word2vec and GloVe, recognized for their ability to capture linguistic regularities. In this paper, we demonstrate a {\em very simple}, and yet counter-intuitive, postprocessing technique -- eliminate the common mean vector and a few top dominating directions from the word vectors -- that renders off-the-shelf representations {\em even stronger}. The postprocessing is empirically validated on a variety of lexical-level intrinsic tasks (word similarity, concept categorization, word analogy) and sentence-level tasks (semantic textural similarity and { text classification}) on multiple datasets and with a variety of representation methods and hyperparameter choices in multiple languages; in each case, the processed representations are consistently better than the original ones.