Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT

Dan Su; Dongyang Dai; Dong Yu; Helen Meng; Jia Jia; Shiyin Kang; Xixin Wu; Zhiyong Wu

REVIEW 5 major objections 5 minor 1 cited by

Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A frozen pre-trained BERT encoder can replace handcrafted feature engineering for Chinese polyphone disambiguation, outperforming a task-trained LSTM baseline.

desk verdict Frozen BERT helps polyphone disambiguation accuracy, but the paper doesn't isolate whether the gain comes from BERT's features or from the per-character output head. read the letter →

arxiv 2501.01102 v1 pith:74KRSFYY submitted 2025-01-02 eess.AS cs.AIcs.SD

Dongyang Dai , Zhiyong Wu , Shiyin Kang , Xixin Wu , Jia Jia , Dan Su , Dong Yu , Helen Meng This is my paper · ORCID

classification eess.AScs.AIcs.SD

keywords polyphonedisambiguationgrapheme-to-phonemeconversionpre-trainedBERTsemanticfeaturestext-to-speechMandarinChineseLSTMbaselineend-to-endframework

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that a frozen, pre-trained BERT encoder can replace handcrafted feature engineering for Chinese polyphone disambiguation, the task of choosing a character's correct pronunciation from context. The authors build an end-to-end system that takes a raw character sequence, lets BERT produce a context-aware semantic feature vector for the polyphonic character, and then a small neural classifier picks the pronunciation. Against the standard LSTM baseline that requires word segmentation and part-of-speech tagging, all three BERT-based variants—with a fully-connected head, an LSTM head, or a Transformer head—report higher cross-validated accuracy. The paper also finds that context closer to the polyphonic character matters more, which explains why the LSTM head beats the Transformer head. If the claim holds, it means large pre-trained language models can absorb the linguistic knowledge needed for Mandarin G2P without task-specific feature engineering.

What carries the argument

The load-bearing mechanism is the frozen pre-trained BERT encoder followed by a per-character unshared output layer. BERT, a multi-layer bidirectional Transformer pre-trained on unlabeled text with masked-character and next-sentence objectives, converts a raw character sequence into a sequence of 768-dimensional semantic vectors; the vector at the polyphonic character's position, optionally refined by an LSTM or Transformer head that models context, is fed to a dedicated softmax output layer unique to that character. The unshared output layer is what lets the system avoid predicting a pronunciation that belongs to a different polyphonic character and accommodates new polyphonic characters by adding an output layer without retraining the shared encoder.

What would settle it

Re-run the same 10-fold cross-validation with the LSTM baseline re-tuned on the development set, and additionally with a strong non-BERT character-level sequence model of similar parameter count. If either closes the accuracy gap to within noise, the claim that pre-trained BERT's semantic features are the source of the improvement is not supported.

Watch

Extended reading notes

Core claim

The central claim is that a pre-trained BERT model, kept frozen during training, extracts sufficiently rich semantic features from a raw Chinese character sequence that a simple downstream classifier can predict polyphonic pronunciations more accurately than a task-trained bidirectional LSTM that uses handcrafted character and part-of-speech embeddings. The paper reports this for three classifier heads (fully-connected, LSTM, Transformer block), all of which outperform the baseline on 10-fold cross-validation. It further claims, based on the relative ordering of the three heads and on averaged attention weights, that nearby context is more influential than distant context for choosing a pronunciation.

Load-bearing premise

The claim of 'greatly enhanced' performance rests on a single baseline whose hyperparameters were copied from the original paper without being re-tuned on the dataset used here; if that baseline is under-powered on this corpus, the observed gap would shrink or disappear under a properly tuned comparison.

Editorial extensions

If this is right

If a frozen pre-trained model is enough, TTS pipelines can drop word segmentation and POS tagging from the G2P front end, simplifying deployment.
The per-character unshared output layer means the system can be extended to new polyphonic characters by training only a new small output head, leaving existing heads intact.
The attention-weight result suggests that future disambiguation models can be built to weight near context heavily, possibly with shorter context windows.
The method can be applied to other Chinese character-disambiguation tasks where pronunciation or reading depends on semantics, such as named-entity readings.
Because BERT is frozen, the classifier is cheap to train, and the semantic features are reusable across many downstream heads.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A direct test the paper does not run: compare the frozen-BERT extractor against the same BERT model with its top layers fine-tuned on the polyphone corpus; the paper notes fine-tuning 'did not achieve desired results' with limited data, but a systematic learning-rate sweep might reverse that conclusion on a larger corpus.
The paper implicitly assumes a single fixed context window for BERT's input; one could test whether truncating the input to a tighter window around the polyphonic character changes accuracy, which would quantify how much distant context actually contributes.
The attention-weight analysis is suggestive but not causal; a controlled perturbation (masking distant vs. nearby characters) would directly test the paper's 'closer context matters more' explanation.
If the central claim generalizes, other frozen pre-trained encoders (e.g., different Transformer variants or multilingual models) should also lift polyphone disambiguation, and a comparison would reveal whether the gain comes from pre-training scale or from the Transformer architecture itself.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper shows that a frozen BERT encoder plus a per-character classifier head gets higher polyphone-disambiguation accuracy than the 2016 BLSTM baseline on 10-fold cross-validation. That is a real result, and the experimental setup is mostly reasonable. What is new is the application of pre-trained BERT to Chinese polyphone disambiguation with three different classifier heads, plus the attention-weight analysis. The end-to-end aspect, with no tokenization or POS tagging, is a practical simplification worth having.

The soft spot is the one the stress-test flags, and it's real: the baseline uses a single shared output layer over all polyphonic characters, while every proposed model gives each character its own output layer. So the accuracy gap could come from the per-character head, which avoids cross-character confusion and encodes character-specific priors, rather than from BERT's semantic representations. The paper itself lists the non-shared output layer as an advantage, so the authors are aware of the change—they just never isolate it. Without an ablation that either gives the baseline per-character heads or gives the BERT features a shared head, the central claim that BERT's semantic features 'greatly enhance' performance is overreach.

Other issues are secondary. The paper reports no numerical accuracies or error bars, only a bar chart. The baseline hyperparameters are copied from the original paper without re-tuning, so the baseline may be under-powered. The dataset is proprietary, and no code is provided, which limits verification. The 2,000-occurrence selection threshold excludes rare characters and should be justified. These are the kind of things a referee can fix.

The paper is worth engaging with: it addresses a concrete problem in Mandarin TTS, the method is simple enough to reimplement, and the confound is instructive. But as it stands, the evidence supports 'a BERT-based system with per-character heads outperforms a shared-head LSTM baseline'—not the stronger mechanistic claim. I'd send it to peer review and ask for the missing ablation and exact numbers. If they can't run the control, the paper should be framed as a system comparison rather than evidence for the semantic-features mechanism.

Referee Report

5 major / 5 minor

Summary. The paper proposes an end-to-end framework for Chinese polyphone disambiguation that feeds a raw Chinese character sequence into a frozen pre-trained BERT model and then into one of three neural classifiers (fully connected, BLSTM, or Transformer block). The classifiers use per-character unshared output layers. The authors compare the three proposed methods against the BLSTM baseline of Shan et al. (2016) on a Tencent TTS corpus using 10-fold cross-validation, and they report in Section 3.4 that all proposed methods outperform the baseline. They also analyze average attention weights and PCA embeddings of BERT features to argue that nearby context is most important for polyphone disambiguation.

Significance. If the central claim is established, the contribution is a simple and practical demonstration that a frozen pre-trained BERT encoder provides useful semantic features for Chinese polyphone disambiguation, with the potential to reduce handcrafted feature engineering. The idea is reasonable and the end-to-end formulation is attractive. However, the current evidence does not isolate the effect of BERT features from the change in output-layer architecture, and the quantitative support is incomplete because no numerical accuracies, deviations, or significance tests are given. The paper has no released code or data, and it does not compare with more recent BERT-based or other neural G2P approaches, so the magnitude of the claimed advance over the state of the art is unclear.

major comments (5)

[Section 3.4, Fig. 6] The central quantitative claim rests entirely on Fig. 6, but the figure is never accompanied by numerical accuracy values, standard deviations over the 10 folds, or any statistical test. The sentence "All of our proposed methods outperforms the LSTM baseline" is therefore not verifiable from the manuscript; the size and reliability of the improvement need to be reported explicitly.
[Sections 1, 2.2.1-2.2.3, 3.2] The comparison between the proposed methods and the baseline is confounded by the output-layer design. The baseline from Shan et al. [7] uses a shared output layer that lists all possible pronunciations, whereas every proposed classifier in Sections 2.2.1-2.2.3 uses a per-character unshared output layer. Consequently, the accuracy gain in Fig. 6 may be caused by the per-character output heads, which add parameters and encode character-specific pronunciation priors, rather than by BERT's semantic features. To support the abstract's claim that "the pre-trained model extracts effective semantic features," the authors should add an architecture-controlled ablation, for example applying the same per-character output heads to non-BERT features, or applying BERT features with the baseline's shared output layer.
[Section 3.2] The LSTM baseline's hyperparameters (BLSTM hidden units 512, two layers, context size 1) are taken directly from [7] without re-tuning on the Tencent corpus. If this baseline is under-tuned or under-powered on the present dataset, the reported gains could overstate the benefit of BERT. The baseline should be tuned on the development set, or at least the authors should report that the settings are competitive on this corpus, perhaps by comparing with a tuned BLSTM on the same data.
[Section 3.4, Fig. 8] The conclusion that "the closer the context information is to the position of the polyphonic character, the more important it is" is not established by the presented evidence. The attention-weight plot is a post-hoc visualization of one Transformer block, and it is not linked quantitatively to disambiguation accuracy. Moreover, the comparison between BERT + LSTM and BERT + Transformer block involves two architectures that differ in many respects, and without error bars or significance testing the claim that LSTM is better because of its ability to model nearby information is not supported.
[Section 3.3] The statement that "fine-tuning the pre-trained BERT model did not achieve desired results" is not accompanied by any experiment, numbers, or training details. Since the decision to freeze BERT is a key design choice, the authors should either report the fine-tuning results or remove the claim.

minor comments (5)

[Abstract and Section 1] There are several typos, including "In out experiments" in the Abstract, "pertained BERT" and "Madarin" in Section 1, and "classifer" in Fig. 4; the manuscript should be carefully proofread.
[Section 3.4] The stray character "中文" appears immediately after "distant locations."; this is likely a formatting artifact and should be removed.
[Section 3.4, Fig. 7] The PCA plots in Fig. 7 are not explained in the text, and it is unclear what conclusion the reader should draw from them; either add an explicit discussion or remove the figure.
[References] Reference [13] (Li et al., 2018) does not appear to be cited anywhere in the body of the paper; please cite it or remove it.
[Section 2.2] The notation "the polyphonic word is the ith element" should be "the polyphonic character," since the input is a character sequence and pronunciations are predicted for individual characters.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: BERT is an external pre-trained artifact, the evaluation is held-out, and no load-bearing argument reduces to a fitted parameter or self-citation.

full rationale

The paper's central claim is that a frozen, externally pre-trained BERT model extracts semantic features that improve Chinese polyphone disambiguation. The evidence is a 10-fold cross-validated accuracy comparison against the LSTM baseline of Shan et al. (2016), which is an independent external reference and not a self-citation. The baseline hyperparameters are explicitly taken from that external paper ('identical to the setting in [7]', Section 3.2), and the BERT model is Google's released checkpoint (Section 3.3), so the feature extractor is not fitted to the test data and its output is not defined in terms of the task labels. The proposed classifiers are trained on held-out training folds with softmax cross-entropy loss, and the reported accuracies are averages over held-out test subsets, so the performance claim is not forced by construction. The attention-weight analysis in Section 3.4 is a post-hoc interpretation of the trained BERT model and is not used to define the predicted labels; therefore it does not introduce circularity. The fact that proposed methods differ from the baseline in output-layer design (per-character unshared layers versus a shared output layer) is a legitimate experimental-confound concern about isolating the effect of BERT features, but it is not a circularity: the claimed result is not equivalent to its own input by definition, and no fitted parameter is renamed as a prediction. No load-bearing argument in the paper reduces to a self-citation or to an author-imported uniqueness theorem.

Assumptions & free parameters 8 free parameters · 3 assumptions · 0 invented entities

The paper is an empirical ML study; the central claim depends on several manually chosen hyperparameters and domain assumptions about transferability of BERT features, but introduces no new theoretical entities.

free parameters (8)

FC hidden units = 512
Chosen for the first fully-connected layer in the FC classifier (Section 3.3).
Dropout rate = 0.5
Applied to the FC classifier during training (Section 3.3).
BLSTM hidden units = 512
For the LSTM-based classifier (Section 3.3).
BLSTM layers = 2
Two-layer BLSTM in the LSTM-based classifier and the baseline (Sections 2.2.2 and 3.2).
Learning rate = 5e-4
Adam optimizer learning rate for FC and LSTM classifiers (Section 3.3).
Transformer dimension = 512
Dimension of Transformer block in the Transformer-based classifier (Section 3.3).
Attention heads = 8
Multi-head attention heads in Transformer classifier, identical to [12] (Section 3.3).
Character selection threshold = >2000 sentences
Polyphonic characters appearing in more than 2,000 sentences were kept, covering 83.7% of samples; excludes rare characters (Section 3.1).

assumptions (3)

domain assumption Semantic context determines polyphone pronunciation
The framework assumes pronunciation can be predicted from BERT semantic features of the raw character sequence (Sections 2.1 and 2.2).
domain assumption Frozen BERT representations transfer to this task
The authors freeze BERT parameters because fine-tuning did not help with limited data, assuming the pre-trained features are already suitable (Section 3.3).
standard math Cross-entropy loss is appropriate for classification
The classifier is trained with softmax cross-entropy loss (Section 2.2.1).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT." pith.science (2026). https://pith.science/paper/74KRSFYY

@misc{pith2026250101102,
  author       = {Pith},
  title        = {Pith review of: Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/74KRSFYY}},
  note         = {Machine review of arXiv:2501.01102}
}

read the original abstract

Grapheme-to-phoneme (G2P) conversion serves as an essential component in Chinese Mandarin text-to-speech (TTS) system, where polyphone disambiguation is the core issue. In this paper, we propose an end-to-end framework to predict the pronunciation of a polyphonic character, which accepts sentence containing polyphonic character as input in the form of Chinese character sequence without the necessity of any preprocessing. The proposed method consists of a pre-trained bidirectional encoder representations from Transformers (BERT) model and a neural network (NN) based classifier. The pre-trained BERT model extracts semantic features from a raw Chinese character sequence and the NN based classifier predicts the polyphonic character's pronunciation according to BERT output. In out experiments, we implemented three classifiers, a fully-connected network based classifier, a long short-term memory (LSTM) network based classifier and a Transformer block based classifier. The experimental results compared with the baseline approach based on LSTM demonstrate that, the pre-trained model extracts effective semantic features, which greatly enhances the performance of polyphone disambiguation. In addition, we also explored the impact of contextual information on polyphone disambiguation.

Figures

Figures reproduced from arXiv: 2501.01102 by the authors.

**Figure 1.** Chinese G2P conversion flow input character is not a polyphonic character, we can directly look up the dictionary to derive its pronunciation. Otherwise, we need a polyphone disambiguation model to predict its pronunciation based on its context information. For Chinese polyphonic characters, their pronunciations are affected by the semantic context information [1] of neighboring characters that may occur before or … view at source ↗

**Figure 2.** Model architecture word tokenization and POS tagging are required in the preprocessing stage in Shan’s approach, greatly reducing the work of feature engineering. However, since this model is trained on limited annotation data, it is difficult to learn enough semantic information for polyphone disambiguation. Besides, all the considered polyphonic characters share the same classifier with only one output layer lis… view at source ↗

**Figure 4.** The NN based classifiers, (a) fully-connected network based classifier, (b) LSTM based classifier, (c) Transformer block based classifier Softmax cross-entropy loss is adopted to train the classifier. The LSTM based classifier and Transformer block based classifier also use the same structure of output layer and loss function. 2.2.2. LSTM based classifer Indicated by [7], contextual information such as the POS of po… view at source ↗

Figures from the paper (3 more)

**Figure 5.** Figure 5: LSTM baseline approach for polyphone disambiguation 3.3. Settings of the proposed approach In our experiments, we adopted the pre-trained BERT model provided by Google to extract semantic features from raw Chi- [PITH_FULL_IMAGE:figures/full_fig_p003_5.png]

**Figure 6.** Figure 6: Experimental results: accuracies of defferent methods, (1) LSTM baseline, (2) BERT + FC, (3) BERT + LSTM, (4) BERT + Transformer block nese character sequence1 , and the detail of BERT is identical to the BERTBASE model described in [11] whose output size of Transform…

**Figure 8.** Figure 8: attention average weight cropped around polyphonic character 4. Conclusions In this paper, we proposed an end-to-end framework for Chinese polyphone disambiguation. The proposed framework accepts raw Chinese character sequence as input without any preprocessing, and i…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Audio Turing Test: Benchmarking the Human-likeness of Large Language Model-based Text-to-Speech Systems in Chinese
cs.SD 2025-05 conditional novelty 6.0 of 10

A new Chinese TTS benchmark and protocol, Audio Turing Test, shows top LLM-based TTS models achieve only about 0.4 out of 1.0 on human-likeness, far below real human speech.

Reference graph

Works this paper leans on

21 extracted references · 16 canonical work pages · cited by 1 Pith paper

[7]

The proposed framework accepts raw Chinese character sequence as input without any prepro- cessing, and it consists of a pre-trained BERT model and a NN based classifier

Conclusions In this paper, we proposed an end-to-end framework for Chinese polyphone disambiguation. The proposed framework accepts raw Chinese character sequence as input without any prepro- cessing, and it consists of a pre-trained BERT model and a NN based classifier. We implemented three classifiers based on neural network and conducted experiments to...

work page
[1]

Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT

Introduction Text-to-speech (TTS) technology has been widely used in voice-assistants, car navigation, e-books and other products. For language based on graphic symbols like Chinese, it is nec- essary to convert the input character sequence into phoneme se- quence before synthesizing speech. Therefore the grapheme-to- phoneme (G2P) conversion component is...

work page Pith review arXiv 2025
[2]

The proposed method predicts pronunciation in an end- to-end way, accepting raw Chinese character sequence containing polyphonic character as input, without the ne- cessity of any preprocessing procedures

work page
[3]

A large amount unsupervised data can be adopted to pre-train the model for extracting semantic information, which will boost the performance of polyphone disam- biguation

work page
[4]

The proposed method uses a non-shared output layer among different polyphonic characters, eliminating the case of mis-predicting to pronunciation of other poly- phonic characters. Furthermore, with this architecture, when new polyphonic characters are required to be pro- cessed, only output layers for these new characters are added and trained without aff...

work page
[5]

The proposed approach The proposed framework consists of a pre-trained BERT and NN based classifier. Depicted in Fig.2, the pertained BERT ex- tracts semantic features from a raw Chinese character sequence containing polyphonic character, the following NN based clas- sifier predicts polyphonic character’s pronunciation according to BERT output. In our res...

work page
[6]

Dataset The experiments were conducted on a dataset extracted from TTS corpus in Tencent AI Lab

Experiment and analysis 3.1. Dataset The experiments were conducted on a dataset extracted from TTS corpus in Tencent AI Lab. There are 331,325 sentences containing polyphonic characters in the corpus. We selected polyphonic characters which appear in more than 2,000 sen- tences accounting for 83.7% of the total polyphonic samples. In our experiments, the...

work page
[8]

The status quo of multi-syllable words and regular method of recognizing,

L. Li, “The status quo of multi-syllable words and regular method of recognizing,” Economic and social development, vol. 8, no. 7, pp. 125–128, 2010

work page 2010

Show all 21 references

[9]

Research and implementation of key techniques of chinese language switching system hj-tts,

D. Zhang, “Research and implementation of key techniques of chinese language switching system hj-tts,” Ph.D. dissertation, Graduate School of the Chinese Academy of Sciences (Institute of Computing Technology), 2000

2000
[10]

Linguistic processing in chi- nese text-to-speech conversion,

L. Cai, H. Wei, and X. Zhou, “Linguistic processing in chi- nese text-to-speech conversion,” Chinese Journal of Information, vol. 9, no. 1, pp. 31–36, 1995

1995
[11]

The broad study of homo- graph disambiguity for mandarin speech synthesis,

W. Wang, S. Hwang, and S. Chen, “The broad study of homo- graph disambiguity for mandarin speech synthesis,” inProceeding of Fourth International Conference on Spoken Language Process- ing. ICSLP’96, vol. 3. IEEE, 1996, pp. 1389–1392

1996
[12]

Maximum entropy based homo- graph disambiguation,

F. Liu, Q. Shi, and J. Tao, “Maximum entropy based homo- graph disambiguation,” in The 9th National Conference on Man- Machine Speech Communication, 2007

2007
[13]

Representation learning: A review and new perspectives,

Y . Bengio, A. Courville, and P. Vincent, “Representation learning: A review and new perspectives,” IEEE transactions on pattern analysis and machine intelligence, vol. 35, no. 8, pp. 1798–1828, 2013

2013
[14]

A bi-directional lstm approach for polyphone disambiguation in mandarin chinese,

C. Shan, L. Xie, and K. Yao, “A bi-directional lstm approach for polyphone disambiguation in mandarin chinese,” in2016 10th In- ternational Symposium on Chinese Spoken Language Processing (ISCSLP). IEEE, 2016, pp. 1–5

2016
[15]

Char2wav: End-to-end speech syn- thesis,

J. Sotelo, S. Mehri, K. Kumar, J. F. Santos, K. Kastner, A. Courville, and Y . Bengio, “Char2wav: End-to-end speech syn- thesis,” 2017

2017
[16]

Tacotron: Towards end-to-end speech synthesis,

Y . Wang, R. Skerry-Ryan, D. Stanton, Y . Wu, R. J. Weiss, N. Jaitly, Z. Yang, Y . Xiao, Z. Chen, S. Bengio et al. , “Tacotron: Towards end-to-end speech synthesis,” arXiv preprint arXiv:1703.10135, 2017

2017 arXiv
[17]

Close to human quality tts with transformer,

N. Li, S. Liu, Y . Liu, S. Zhao, M. Liu, and M. Zhou, “Close to human quality tts with transformer,” arXiv preprint arXiv:1809.08895, 2018

2018 arXiv
[18]

Bert: Pre- training of deep bidirectional transformers for language under- standing,

J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre- training of deep bidirectional transformers for language under- standing,” arXiv preprint arXiv:1810.04805, 2018

2018 arXiv
[19]

Attention is all you need,

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” in Advances in Neural Information Processing Systems, 2017, pp. 5998–6008

2017
[20]

Analogical rea- soning on chinese morphological and semantic relations,

S. Li, Z. Zhao, R. Hu, W. Li, T. Liu, and X. Du, “Analogical rea- soning on chinese morphological and semantic relations,” arXiv preprint arXiv:1805.06504, 2018

2018 arXiv
[21]

Adam: A method for stochastic opti- mization,

D. P. Kingma and J. Ba, “Adam: A method for stochastic opti- mization,” arXiv preprint arXiv:1412.6980, 2014

2014 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [7]

The proposed framework accepts raw Chinese character sequence as input without any prepro- cessing, and it consists of a pre-trained BERT model and a NN based classifier

Conclusions In this paper, we proposed an end-to-end framework for Chinese polyphone disambiguation. The proposed framework accepts raw Chinese character sequence as input without any prepro- cessing, and it consists of a pre-trained BERT model and a NN based classifier. We implemented three classifiers based on neural network and conducted experiments to...

work page

[2] [1]

Disambiguation of Chinese Polyphones in an End-to-End Framework with Semantic Features Extracted by Pre-trained BERT

Introduction Text-to-speech (TTS) technology has been widely used in voice-assistants, car navigation, e-books and other products. For language based on graphic symbols like Chinese, it is nec- essary to convert the input character sequence into phoneme se- quence before synthesizing speech. Therefore the grapheme-to- phoneme (G2P) conversion component is...

work page Pith review arXiv 2025

[3] [2]

The proposed method predicts pronunciation in an end- to-end way, accepting raw Chinese character sequence containing polyphonic character as input, without the ne- cessity of any preprocessing procedures

work page

[4] [3]

A large amount unsupervised data can be adopted to pre-train the model for extracting semantic information, which will boost the performance of polyphone disam- biguation

work page

[5] [4]

The proposed method uses a non-shared output layer among different polyphonic characters, eliminating the case of mis-predicting to pronunciation of other poly- phonic characters. Furthermore, with this architecture, when new polyphonic characters are required to be pro- cessed, only output layers for these new characters are added and trained without aff...

work page

[6] [5]

The proposed approach The proposed framework consists of a pre-trained BERT and NN based classifier. Depicted in Fig.2, the pertained BERT ex- tracts semantic features from a raw Chinese character sequence containing polyphonic character, the following NN based clas- sifier predicts polyphonic character’s pronunciation according to BERT output. In our res...

work page

[7] [6]

Dataset The experiments were conducted on a dataset extracted from TTS corpus in Tencent AI Lab

Experiment and analysis 3.1. Dataset The experiments were conducted on a dataset extracted from TTS corpus in Tencent AI Lab. There are 331,325 sentences containing polyphonic characters in the corpus. We selected polyphonic characters which appear in more than 2,000 sen- tences accounting for 83.7% of the total polyphonic samples. In our experiments, the...

work page

[8] [8]

The status quo of multi-syllable words and regular method of recognizing,

L. Li, “The status quo of multi-syllable words and regular method of recognizing,” Economic and social development, vol. 8, no. 7, pp. 125–128, 2010

work page 2010