Multilingual Grapheme-To-Phoneme Conversion with Byte Representation

Yu, Mingzhi; Nguyen, Hieu Duy; Sokolov, Alex; Lepird, Jack; Sathyendra, Kanthashree Mysore; Choudhary, Samridhi; Mouchtaris, Athanasios; Kunzmann, Siegfried

doi:10.1109/icassp40776.2020.9054696

Cited by 10 publications

(15 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lexicon texts have been used in word sense disambiguation [13] and machine translation (MT) [14], while we apply the idea to TTS. The particular task we apply to is closely related to G2P with rich prior work, particularly on Chinese and Japanese [15,16], possibly enhanced by implicit external knowledge [17], and lexicons have been used during training [18,19]. While the flexibility of using external knowledge has been discussed in MT [20] and G2P [21], though knowledge from raw lexicon texts were not leveraged.…”

Section: Related Workmentioning

confidence: 99%

Neural Lexicon Reader: Reduce Pronunciation Errors in End-to-end TTS by Leveraging External Textual Knowledge

He¹,

Yang²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

End-to-end TTS suffers from high data requirements as it is difficult for both costly speech corpora to cover all necessary knowledge and neural models to learn the knowledge, hence additional knowledge needs to be injected manually. For example, to capture pronunciation knowledge on languages without regular orthography, a complicated grapheme-tophoneme pipeline needs to be built based on a structured, large pronunciation lexicon, leading to extra, sometimes high, costs to extend neural TTS to such languages. In this paper, we propose a framework to learn to extract knowledge from unstructured external resources using Token2Knowledge attention modules. The framework is applied to build a novel end-to-end TTS model named Neural Lexicon Reader that extracts pronunciations from raw lexicon texts. Experiments support the potential of our framework that the model significantly reduces pronunciation errors in low-resource, end-toend Chinese TTS, and the lexicon-reading capability can be transferred to other languages with a smaller amount of data.

show abstract

Section: Related Workmentioning

confidence: 99%

Neural Lexicon Reader: Reduce Pronunciation Errors in End-to-end TTS by Leveraging External Textual Knowledge

He¹,

Yang²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Even for the mainstream languages, large, manually annotated lexicons are not easily and readily accessible. And most research groups have developed their internal resources [9,14]. An alternative to this individual effort is the collaborative online resource called Wiktionary.…”

Section: Phonetic Lexiconsmentioning

confidence: 99%

“…As a result, more and more NLP tasks have been addressed with CNN or hybrid seq2seq alternatives [20,21]. Along the CNN-based architecture, the Transformer network has been successfully applied in machine translation tasks [22], and G2P conversion networks [9,23].…”

Section: G2p Conversion Networkmentioning

confidence: 99%

“…For these two neural architectures, the topologies which obtained the best results for English are described in [9,23,24]. However, taking into consideration the G2P complexity across languages, as well as the variable dimension of each phonetic lexicon, the architectures' hyperparameters need to be optimized [25].…”

Section: G2p Conversion Networkmentioning

confidence: 99%

“…The authors of [8] also use several online repositories to train and adapt the models from high-resource languages to related low-resource languages. Multilingual G2P was also addressed by changing the grapheme representation: [9] proposes a model which uses byte-level input representation to accommodate different grapheme systems, along with an attention-based Transformer architecture. Ancillary audio data was also used to learn a more optimal intermediate representation of source graphemes in a multi-task training process for multilingual G2P [10].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

RECOApy: Data Recording, Pre-Processing and Phonetic Transcription for End-to-End Speech-Based Applications

Stan¹

2020

Interspeech 2020

View full text Add to dashboard Cite

Deep learning enables the development of efficient end-to-end speech processing applications while bypassing the need for expert linguistic and signal processing features. Yet, recent studies show that good quality speech resources and phonetic transcription of the training data can enhance the results of these applications. In this paper, the RECOApy tool is introduced. RECOApy streamlines the steps of data recording and pre-processing required in end-to-end speech-based applications. The tool implements an easy-to-use interface for prompted speech recording, spectrogram and waveform analysis, utterance-level normalisation and silence trimming, as well grapheme-to-phoneme conversion of the prompts in eight languages: Czech, English, French, German, Italian, Polish, Romanian and Spanish.The grapheme-to-phoneme (G2P) converters are deep neural network (DNN) based architectures trained on lexicons extracted from the Wiktionary online collaborative resource. With the different degree of orthographic transparency, as well as the varying amount of phonetic entries across the languages, the DNN's hyperparameters are optimised with an evolution strategy. The phoneme and word error rates of the resulting G2P converters are presented and discussed. The tool, the processed phonetic lexicons and trained G2P models are made freely available.

show abstract

Literature Review of Research on Common Methods of Grapheme-To-Phoneme

Zhang,

Cao

2023

Ieis 2022

View full text Add to dashboard Cite

Multilingual Grapheme-To-Phoneme Conversion with Byte Representation

Cited by 10 publications

References 11 publications

Neural Lexicon Reader: Reduce Pronunciation Errors in End-to-end TTS by Leveraging External Textual Knowledge

Neural Lexicon Reader: Reduce Pronunciation Errors in End-to-end TTS by Leveraging External Textual Knowledge

RECOApy: Data Recording, Pre-Processing and Phonetic Transcription for End-to-End Speech-Based Applications

Literature Review of Research on Common Methods of Grapheme-To-Phoneme

Contact Info

Product

Resources

About