Spell Once, Summon Anywhere: A Two-Level Open-Vocabulary Language Model

TL;DR

This paper introduces a two-level neural language model that leverages word spellings to handle unknown words, improving open-vocabulary NLP performance by combining sentence and word structure modeling.

Contribution

It proposes a novel Bayesian generative framework integrating RNN-based sentence and spelling models for open-vocabulary language modeling.

Findings

Achieves state-of-the-art results on multiple datasets.

Effectively models unknown words using spelling information.

Outperforms previous methods including a strong baseline.

Abstract

We show how the spellings of known words can help us deal with unknown words in open-vocabulary NLP tasks. The method we propose can be used to extend any closed-vocabulary generative model, but in this paper we specifically consider the case of neural language modeling. Our Bayesian generative story combines a standard RNN language model (generating the word tokens in each sentence) with an RNN-based spelling model (generating the letters in each word type). These two RNNs respectively capture sentence structure and word structure, and are kept separate as in linguistics. By invoking the second RNN to generate spellings for novel words in context, we obtain an open-vocabulary language model. For known words, embeddings are naturally inferred by combining evidence from type spelling and token context. Comparing to baselines (including a novel strong baseline), we beat previous work and establish state-of-the-art results on multiple datasets.