FRAGE: Frequency-Agnostic Word Representation

TL;DR

This paper introduces FRAGE, a frequency-agnostic word embedding method using adversarial training, which improves the effectiveness of word representations, especially for rare words, across multiple NLP tasks.

Contribution

The paper proposes a novel adversarial training approach to learn frequency-agnostic word embeddings, addressing bias towards word frequency in existing embeddings.

Findings

FRAGE outperforms baselines in word similarity tasks

FRAGE improves performance in language modeling and machine translation

FRAGE enhances text classification results

Abstract

Continuous word representation (aka word embedding) is a basic building block in many neural network-based models used in natural language processing tasks. Although it is widely accepted that words with similar semantics should be close to each other in the embedding space, we find that word embeddings learned in several tasks are biased towards word frequency: the embeddings of high-frequency and low-frequency words lie in different subregions of the embedding space, and the embedding of a rare word and a popular word can be far from each other even if they are semantically similar. This makes learned word embeddings ineffective, especially for rare words, and consequently limits the performance of these neural network models. In this paper, we develop a neat, simple yet effective way to learn \emph{FRequency-AGnostic word Embedding} (FRAGE) using adversarial training. We conducted comprehensive studies on ten datasets across four natural language processing tasks, including word similarity, language modeling, machine translation and text classification. Results show that with FRAGE, we achieve higher performance than the baselines in all tasks.