G2P 中文转拼音多音字消歧论文总结

一时失言乱红尘 2024-04-03 14:14 126阅读 0赞

***论文链接均在作者Github下，大家多多专注呀！***

## 1. 字形转音素 ##

### 1.1 基于规则 ###

(1998) [Issues in building general letter to sound rules][]

### 1.2 基于联合序列模型 ###

(2002) [Pronunciation of proper names with a joint n-gram model for bi-directional grapheme-to-phoneme conversion][]  
(2008) [Joint-sequence models for grapheme-to-phoneme conversion][]  
(ICASSP 2017) [Bayesian\_joint-sequence\_models\_for\_grapheme-to-phoneme\_conversion][Bayesian_joint-sequence_models_for_grapheme-to-phoneme_conversion]

### 1.3 基于编码-解码器结构 ###

(Interspeech 2015) [Sequence-to-sequence neural net models for grapheme-to-phoneme conversion][]  
(Interspeech 2017) [Multitask sequence-to-sequence models for grapheme-to-phoneme conversion][]  
(ICASSP 2018) [Convolutional sequence to sequence model with non-sequential greedy decoding for grapheme to phoneme conversion][]  
(2019) [Grapheme-to-phoneme conversion with convolutional neural networks][]  
(ICASSP 2022) [Neural\_Grapheme-To-Phoneme\_Conversion\_with\_Pre-Trained\_Grapheme\_Models][Neural_Grapheme-To-Phoneme_Conversion_with_Pre-Trained_Grapheme_Models]  
(ICASSP 2021) [Joint\_Alignment\_Learning-Attention\_Based\_Model\_for\_Grapheme-to-Phoneme\_Conversion][Joint_Alignment_Learning-Attention_Based_Model_for_Grapheme-to-Phoneme_Conversion]

### 1.4 基于长短时记忆循环神经网络 ###

(ICASSP 2015) [GRAPHEME-TO-PHONEME CONVERSION USING LONG SHORT-TERM MEMORY RECURRENT NEURAL NETWORKS][]

### 1.5 基于多语言编码 ###

(2016) [Grapheme-to-phoneme models for any language][]  
(2017) [Massively multilingual neural grapheme-to-phoneme conversion][Grapheme-to-phoneme models for any language]  
(2018) [Epitran: Precision g2p for many languages][Epitran_ Precision g2p for many languages]

### 1.6 基于小样本的循环神经网络模型 ###

(ICASSP 2017) [Low-resource\_grapheme-to-phoneme\_conversion\_using\_recurrent\_neural\_networks][Low-resource_grapheme-to-phoneme_conversion_using_recurrent_neural_networks]

## 2. 中文转拼音 ##

### 2.1 基于规则 ###

(1998) [Chinese sentence tokenization using Viterbi decoder][]  
(2004) [Grapheme-to-phoneme conversion for Chinese text-to-speech][]  
(2005) [Grapheme-to-phoneme\_conversion\_in\_Chinese\_TTS\_system][Grapheme-to-phoneme_conversion_in_Chinese_TTS_system]  
(2006) [汉语字音转换中的多层面多音字读音消歧][Link 1]  
(2008) [中文TTS系统中多音字的一种解决方案][Link 1]  
(2008)[Disambiguating\_effectively\_Chinese\_polyphonic\_ambiguity\_based\_on\_unify\_approach][Disambiguating_effectively_Chinese_polyphonic_ambiguity_based_on_unify_approach]  
(2010) [Disambiguation for polyphones of Chinese based on two-pass unified approach][]

### 2.2 基于机器学习 ###

#### 2.2.1 基于决策树分类 ####

(1996) [Comparison of Two Tree-Structured Approaches for Grapheme-to-Phoneme Conversion][]  
(1996) [The broad study of homograph disambiguity for mandarin speech synthesis][]  
(2002) [An efficient way to learn rules for grapheme-to-phoneme conversion in Chinese][]  
(2010) [Polyphonic word disambiguation with machine learning approaches][]

#### 2.2.2 基于最大熵 ####

(2007) [基于最大熵模型的多音字消歧][Link 2]  
(2007) [Inequality maximum entropy classifier with character features for polyphone disambiguation in mandarin tts systems][]  
(2011) [Polyphone disambiguation based on maximum entropy model in mandarin grapheme-to-phoneme conversion][]

#### 2.2.4 基于TBL算法 ####

(2005) [Grapheme-to-phoneme conversion based on TBL algorithm in Mandarin TTS systems][]  
(ICASSP2008) [Tree-guided transformation-based homograph disambiguation in Mandarin TTS system][]

### 2.3 基于深度学习 ###

#### 2.3.1 基于LSTM ####

(ICASSP 2016) [A Bi-directional LSTM Approach for Polyphone Disambiguation in Mandarin Chinese][]  
(2020) [基于深度学习的中文至拼音首字母自动转化方法][Link 3]  
(2021)[A Method of Polyphone Disambiguation Based on Semantic Extension][]

#### 2.3.2 基于神经网络 ####

(2019) [Polyphone Disambiguation for Mandarin Chinese Using Conditional Neural Network with Multi-level Embedding Features][]  
(Interspeech 2020) [A Mask-based Model for Mandarin Chinese Polyphone Disambiguation][]

#### 2.3.3 基于Encoder\_Decoder结构 ####

(Interspeech 2019) [Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-Trained BERT][]  
(2020) [Multilingual grapheme-to-phoneme conversion with byte representation][]  
(2020) [Neural Machine Translation for Multilingual Grapheme-to-Phoneme Conversion][]

#### 2.3.4 基于预训练模型 ####

(Interspeech 2019) [Pre-Trained Text Representations for Improving Front-End Text Processing in Mandarin Text-to-Speech Synthesis][]  
(ASRU 2019) [Knowledge\_Distillation\_from\_Bert\_in\_Pre-Training\_and\_Fine-Tuning\_for\_Polyphone\_Disambiguation][Knowledge_Distillation_from_Bert_in_Pre-Training_and_Fine-Tuning_for_Polyphone_Disambiguation]  
(ICASSP 2021) [A Universal Bert-Based Front-End Model for Mandarin Text-To-Speech Synthesis][]  
(ICASSP 2022) [Data\_Augmentation\_for\_Long-Tailed\_and\_Imbalanced\_Polyphone\_Disambiguation\_in\_Mandarin][A Universal Bert-Based Front-End Model for Mandarin Text-To-Speech Synthesis]

### 2.4 基于新工具包 ###

(2020): [g2pM：A Neural Grapheme-to-Phoneme Conversion Package for Mandarin Chinese Based on a New Open Benchmark Dataset][g2pM_A Neural Grapheme-to-Phoneme Conversion Package for Mandarin Chinese Based on a New Open Benchmark Dataset]

[Issues in building general letter to sound rules]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Issues%20in%20building%20general%20letter%20to%20sound%20rules.pdf
[Pronunciation of proper names with a joint n-gram model for bi-directional grapheme-to-phoneme conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Pronunciation%20of%20proper%20names%20with%20a%20joint%20n-gram%20model%20for%20bi-directional%20grapheme-to-phoneme%20conversion.pdf
[Joint-sequence models for grapheme-to-phoneme conversion]: https://www.sciencedirect.com/science/article/abs/pii/S0167639308000046
[Bayesian_joint-sequence_models_for_grapheme-to-phoneme_conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2017%20ICASSP%20Bayesian_joint-sequence_models_for_grapheme-to-phoneme_conversion.pdf
[Sequence-to-sequence neural net models for grapheme-to-phoneme conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Sequence-to-sequence%20neural%20net%20models%20for%20grapheme-to-phoneme%20conversion.pdf
[Multitask sequence-to-sequence models for grapheme-to-phoneme conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Multitask%20sequence-to-sequence%20models%20for%20grapheme-to-phoneme%20conversion.PDF
[Convolutional sequence to sequence model with non-sequential greedy decoding for grapheme to phoneme conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2018%20ICASSP%20Convolutional_Sequence_to_Sequence_Model_with_Non-Sequential_Greedy_Decoding_for_Grapheme_to_Phoneme_Conversion.pdf
[Grapheme-to-phoneme conversion with convolutional neural networks]: https://www.mdpi.com/2076-3417/9/6/1143
[Neural_Grapheme-To-Phoneme_Conversion_with_Pre-Trained_Grapheme_Models]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2022%20ICASSP%20Neural_Grapheme-To-Phoneme_Conversion_with_Pre-Trained_Grapheme_Models.pdf
[Joint_Alignment_Learning-Attention_Based_Model_for_Grapheme-to-Phoneme_Conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2021%20ICASSP%20Joint_Alignment_Learning-Attention_Based_Model_for_Grapheme-to-Phoneme_Conversion.pdf
[GRAPHEME-TO-PHONEME CONVERSION USING LONG SHORT-TERM MEMORY RECURRENT NEURAL NETWORKS]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2015%20ICASSP%20Grapheme-to-phoneme_conversion_using_Long_Short-Term_Memory_recurrent_neural_networks.pdf
[Grapheme-to-phoneme models for any language]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Grapheme-to-phoneme%20models%20for%20%28Almost%29%20Any%20language.pdf
[Epitran_ Precision g2p for many languages]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Epitran%20Precision%20g2p%20for%20many%20languages.pdf
[Low-resource_grapheme-to-phoneme_conversion_using_recurrent_neural_networks]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2017%20ICASSP%20Low-resource_grapheme-to-phoneme_conversion_using_recurrent_neural_networks.pdf
[Chinese sentence tokenization using Viterbi decoder]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Chinese%20sentence%20tokenization%20using%20viterbi%20decoder.pdf
[Grapheme-to-phoneme conversion for Chinese text-to-speech]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Grapheme-to-Phoneme%20Conversion%20for%20Chinese%20Text-to-Speech.pdf
[Grapheme-to-phoneme_conversion_in_Chinese_TTS_system]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/2004%20Grapheme-to-phoneme_conversion_in_Chinese_TTS_system.pdf
[Link 1]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/%E6%B1%89%E8%AF%AD%E5%AD%97%E9%9F%B3%E8%BD%AC%E6%8D%A2%E4%B8%AD%E7%9A%84%E5%A4%9A%E5%B1%82%E9%9D%A2%E5%A4%9A%E9%9F%B3%E5%AD%97%E8%AF%BB%E9%9F%B3%E6%B6%88%E6%AD%A7_%E8%8C%83%E6%98%8E.caj
[Disambiguating_effectively_Chinese_polyphonic_ambiguity_based_on_unify_approach]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Disambiguating_effectively_Chinese_polyphonic_ambiguity_based_on_unify_approach.pdf
[Disambiguation for polyphones of Chinese based on two-pass unified approach]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Disambiguation_for_polyphones_of_Chinese_based_on_two-pass_unified_approach.pdf
[Comparison of Two Tree-Structured Approaches for Grapheme-to-Phoneme Conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Comparison_of_two_tree-structured_approaches_for_grapheme-to-phoneme_conversion.pdf
[The broad study of homograph disambiguity for mandarin speech synthesis]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/The_broad_study_of_homograph_disambiguity_for_Mandarin_speech_synthesis.pdf
[An efficient way to learn rules for grapheme-to-phoneme conversion in Chinese]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/An%20efficient%20way%20to%20learn%20rules%20for%20grapheme-to-phoneme%20conversion%20in%20Chinese.pdf
[Polyphonic word disambiguation with machine learning approaches]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Polyphonic_Word_Disambiguation_with_Machine_Learning_Approaches.pdf
[Link 2]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/%E5%9F%BA%E4%BA%8E%E6%9C%80%E5%A4%A7%E7%86%B5%E6%A8%A1%E5%9E%8B%E7%9A%84%E5%A4%9A%E9%9F%B3%E5%AD%97%E6%B6%88%E6%AD%A7_%E5%88%98%E6%96%B9%E8%88%9F.pdf
[Inequality maximum entropy classifier with character features for polyphone disambiguation in mandarin tts systems]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Inequality_Maximum_Entropy_Classifier_with_Character_Features_for_Polyphone_Disambiguation_in_Mandarin_TTS_Systems.pdf
[Polyphone disambiguation based on maximum entropy model in mandarin grapheme-to-phoneme conversion]: https://www.scientific.net/KEM.480-481.1043
[Grapheme-to-phoneme conversion based on TBL algorithm in Mandarin TTS systems]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Grapheme-to-phoneme%20conversion%20based%20on%20TBL%20algorithm%20in.pdf
[Tree-guided transformation-based homograph disambiguation in Mandarin TTS system]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Tree-guided_transformation-based_homograph_disambiguation_in_Mandarin_TTS_system.pdf
[A Bi-directional LSTM Approach for Polyphone Disambiguation in Mandarin Chinese]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/A%20Bi-directional%20LSTM%20Approach%20for%20Polyphone%20Disambiguation%20in%20Mandarin%20chinese.pdf
[Link 3]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/%E5%9F%BA%E4%BA%8E%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%9A%84%E4%B8%AD%E6%96%87%E8%87%B3%E6%8B%BC%E9%9F%B3%E9%A6%96%E5%AD%97%E6%AF%8D%E8%87%AA%E5%8A%A8%E8%BD%AC%E5%8C%96%E6%96%B9%E6%B3%95_%E8%83%A1%E5%8D%87%E6%B3%BD.pdf
[A Method of Polyphone Disambiguation Based on Semantic Extension]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/A_Method_of_Polyphone_Disambiguation_Based_on_Semantic_Extension.pdf
[Polyphone Disambiguation for Mandarin Chinese Using Conditional Neural Network with Multi-level Embedding Features]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Polyphone%20Disambiguation%20for%20Mandarin%20Chinese%20Using.pdf
[A Mask-based Model for Mandarin Chinese Polyphone Disambiguation]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/A%20Mask-based%20Model%20for%20Mandarin%20Chinese%20Polyphone%20Disambiguation.pdf
[Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-Trained BERT]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Disambiguation%20of%20Chinese%20Polyphones%20in%20an%20End-to-End.pdf
[Multilingual grapheme-to-phoneme conversion with byte representation]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Multilingual_Grapheme-To-Phoneme_Conversion_with_Byte_Representation.pdf
[Neural Machine Translation for Multilingual Grapheme-to-Phoneme Conversion]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Neural%20Machine%20Translation%20for%20Multilingual%20Grapheme-to-Phoneme.pdf
[Pre-Trained Text Representations for Improving Front-End Text Processing in Mandarin Text-to-Speech Synthesis]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Pre-Trained%20Text%20Representations.pdf
[Knowledge_Distillation_from_Bert_in_Pre-Training_and_Fine-Tuning_for_Polyphone_Disambiguation]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/Knowledge_Distillation_from_Bert_in_Pre-Training_and_Fine-Tuning_for_Polyphone_Disambiguation.pdf
[A Universal Bert-Based Front-End Model for Mandarin Text-To-Speech Synthesis]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/A_Universal_Bert-Based_Front-End_Model_for_Mandarin_Text-To-Speech_Synthesis.pdf
[g2pM_A Neural Grapheme-to-Phoneme Conversion Package for Mandarin Chinese Based on a New Open Benchmark Dataset]: https://github.com/luodeyu9990714/Papers-for-G2P-and-ChineseG2P/blob/main/g2pM%EF%BC%9AA%20Neural%20Grapheme-to-Phoneme%20Conversion%20Package%20for%20Mandarin%20Chinese%20Based%20on%20a%20New%20Open%20Benchmark%20Dataset.pdf