Who Needs Words? Lexicon-Free Speech Recognition

Citations Over TimeTop 10% of 2019 papers

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.

Related Papers

• → The evolving nature of the Japanese lexicon in a tradition of lesson study(2021)5 cited
• → Automatic Construction of a Core Lexicon for Specific Domain(2007)6 cited
• VARIAN LEKSIKON BAHASA JAWA PADA WILAYAH SEGITIGA DI KABUPATEN WONOSOBO(2017)
• → SELF & FEIL: Emotion and Intensity Lexicons for Finnish(2021)1 cited
• An Analysis of Combining and Pragmatic Motivations on "Internet Fashionable Lexicon(2003)