Generative Spoken Language Model based on continuous word-sized audio tokens - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

Generative Spoken Language Model based on continuous word-sized audio tokens

Abstract

In NLP, text language models based on words or subwords are known to outperform their character-based counterparts. Yet, in the speech community, the standard input of spoken LMs are 20ms or 40ms-long discrete units (shorter than a phoneme). Taking inspiration from wordbased LM, we introduce a Generative Spoken Language Model (GSLM) based on word-size continuous-valued audio embeddings that can generate diverse and expressive language output. This is obtained by replacing lookup table for lexical types with a Lexical Embedding function, the cross entropy loss by a contrastive loss, and multinomial sampling by k-NN sampling. The resulting model is the first generative language model based on word-size continuous embeddings. Its performance is on par with discrete unit GSLMs regarding generation quality as measured by automatic metrics and subjective human judgements. Moreover, it is five times more memory efficient thanks to its large 200ms units. In addition, the embeddings before and after the Lexical Embedder are phonetically and semantically interpretable.
Fichier principal
Vignette du fichier
tGSLM__Robin_A___EMNLP_2023.pdf (724.89 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-04402373 , version 1 (18-01-2024)

Licence

Attribution

Identifiers

  • HAL Id : hal-04402373 , version 1

Cite

Robin Algayres, Yossi Adi, Tu Anh Nguyen, Jade Copet, Gabriel Synnaeve, et al.. Generative Spoken Language Model based on continuous word-sized audio tokens. The 2023 Conference on Empirical Methods in Natural Language Processing, Dec 2023, Singapore, Singapore. ⟨hal-04402373⟩
16 View
24 Download

Share

Gmail Facebook X LinkedIn More