Deep contextualized word representations

TL;DR

This paper introduces deep contextualized word representations derived from a pre-trained bidirectional language model, significantly enhancing NLP task performance by capturing complex word usage and contextual variations.

Contribution

It presents a novel deep contextualized word embedding method based on biLMs, improving NLP tasks and providing insights into internal network representations.

Findings

Improved performance across six NLP tasks.

Deep internals of the model are crucial for effectiveness.

Representations model both syntax and semantics.

Abstract

We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.