# Who Needs Words? Lexicon-Free Speech Recognition

## TL;DR

This paper demonstrates that lexicon-free, character-based language models can match or outperform traditional lexicon-based models in speech recognition, especially for out-of-vocabulary words, by leveraging large context through convolutional architectures.

## Contribution

It shows that character-based LMs, particularly convolutional ones, can effectively replace word-based LMs in lexicon-free speech recognition, improving OOV word handling.

## Key findings

- Character-based LMs perform as well as word-based LMs in WER.
- Convolutional LMs effectively leverage large contexts.
- Lexicon-free decoding outperforms lexicon-based methods on OOV words.

## Abstract

Lexicon-free speech recognition naturally deals with the problem of out-of-vocabulary (OOV) words. In this paper, we show that character-based language models (LM) can perform as well as word-based LMs for speech recognition, in word error rates (WER), even without restricting the decoding to a lexicon. We study character-based LMs and show that convolutional LMs can effectively leverage large (character) contexts, which is key for good speech recognition performance downstream. We specifically show that the lexicon-free decoding performance (WER) on utterances with OOV words using character-based LMs is better than lexicon-based decoding, both with character or word-based LMs.