Pith. sign in

REVIEW 5 major objections 6 minor 26 references

Automatic Text Pronunciation Correlation Generation and Application for Contextual Biasing

T0 review · 5 major / 6 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 This paper proposes ATPC, a data-driven method that builds a character-level pronunciation-distance matrix from speech and text alone, and shows it improves contextual biasing in Mandarin end-to-end ASR without a manual lexicon.

desk verdict The core biasing result is real, but the lexicon-free selling point collapses on reading Section III-C: the speech encoder is fine-tuned on IPA derived from a hand-built BABEL lexicon, so ATPC still leans on a manually crafted lexicon. read the letter →

arxiv 2501.00804 v1 pith:V46YJIJD submitted 2025-01-01 eess.AS cs.CL

Gaofeng Cheng , Haitian Lu , Chengxu Yang , Xuyang Wang , Ta Li , Yonghong Yan This is my paper · ORCID
classification eess.AScs.CL
keywords automatictextpronunciationcorrelationcontextualbiasingend-to-endASRlexicondynamictimewarpingspeechembeddingsMandarintimestampalignment
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 proposes a fully data-driven way to learn pronunciation correlations between written characters and shows that the learned correlations improve hotword-biased Mandarin speech recognition. The method, called automatic text pronunciation correlation (ATPC), needs only the same speech-plus-transcript supervision already used to train end-to-end ASR systems, and it produces a pairwise distance matrix between characters without any hand-built pronunciation lexicon. In experiments, ATPC lowers character error rate by 13.0% relative and biased-character error rate by 22.5% relative over the no-biasing baseline, and it outperforms a separately trained deep-biasing network while leaving the original ASR architecture untouched. A sympathetic reading of the paper is that it establishes ATPC as a viable data-driven substitute for pronunciation lexicons in contextual biasing, at least for Mandarin.

What carries the argument

The central object is the ATPC matrix, a table of pairwise pronunciation distances between Mandarin characters. It is built in three stages: the iteratively-trained timestamp estimator (ITSE) produces start and end timestamps for every character in each training utterance; a multilingual speech encoder fine-tuned on IPA recognition extracts frame-level speech embeddings at 50 Hz; and dynamic time warping with cosine distance averages segment-to-segment distances over many embeddings of each character, giving one distance per character pair. At decoding time the matrix is row-normalized with diagonal set to 1.0, characters closer than a threshold of 1.07 are treated as candidate replacements for recognized characters, and hotword replacements are applied in ascending distance order.

What would settle it

Take an ATPC matrix built on a source corpus, apply it to a held-out target corpus from different speakers or recording conditions, and measure the DTW distance gap between known homophone pairs and known non-homophone pairs; if the gap disappears or reverses, or if substituting the matrix into the biasing decoder no longer lowers CER, the cross-corpus transfer claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that text-to-text pronunciation correlation can be extracted automatically from audio and text by aligning each character to its speech segment, embedding those segments with a multilingual speech representation model, and measuring pairwise dynamic-time-warping distances between characters. These distances make up the ATPC matrix, and the paper shows that inside the matrix homophones sit close together while unrelated characters sit far apart. Plugging the matrix into an end-to-end ASR decoder as a hotword biasing module yields a relative CER reduction of 13.0% and a relative B-CER reduction of 22.5% over the baseline, with hotword recall up 25 percentage points and F1 up 24 percentage points, all without training any additional neural network.

Load-bearing premise

The whole comparison rests on the assumption that pronunciation distances measured on one Mandarin corpus stay about the same on a different Mandarin corpus, so a single cutoff of 1.07 still marks confusable characters in the test domain.

Editorial extensions

If this is right

  • Contextual biasing for Mandarin can be improved from speech-transcript pairs alone, with no extra neural network to train and no manual lexicon.
  • ATPC combines with a WFST context graph, and the combined system reaches 10.3% CER and an F1 of 80 on the hotword test set, beating either module alone.
  • The same generation pipeline can in principle be applied to dialects or languages that lack pronunciation lexicons, which is the paper's stated motivation.
  • The method still trails hand-built lexicons on the same test set, so its practical role is to fill gaps where such lexicons do not exist.

Reading between the lines

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

  • Beyond the paper's experiments, the same matrix could serve as a pseudo-pronunciation lexicon for text-to-speech or pronunciation-aware spelling correction, since it encodes phonetic similarity without phonetic labels.
  • The distance threshold 1.07 was tuned on a development set; porting ATPC to another domain will likely require re-estimating that threshold, and the reported gains may not survive an unchanged threshold.
  • The paper does not directly test how well an ATPC matrix transfers across corpora; measuring the homophone/non-homophone distance gap on a held-out domain would quantify that transfer.
  • As the matrix is derived from data, it can track named entities and new vocabulary as transcripts accumulate, though characters never seen in training remain an open problem.
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.

Referee Report

5 major / 6 minor

Summary. The paper proposes a data-driven method, ATPC, to automatically derive a character-level pronunciation-distance matrix from speech and text annotations, and applies it to contextual biasing in end-to-end ASR. The pipeline first uses the authors' ITSE algorithm to align text symbols with speech, then extracts speech embeddings from XLSR-53 (including a variant fine-tuned on BABEL IPA recognition), and finally computes DTW distances between embedding sets to fill a 3711 x 3711 matrix of Mandarin characters. The matrix is used in a WFST-based contextual biasing decoder to replace ASR output characters with nearby hotword characters. On the Aishell-1 contextual biasing test set, the authors report that ATPC reduces CER from 13.8% to 12.0% (a 13.0% relative reduction) and B-CER by 22.5% relative, while C-g+ATPC reaches 10.3% CER, approaching the C-g+Manual Lexicons result of 8.9% CER. The paper concludes that ATPC is a promising lexicon-free approach for languages and dialects lacking manually crafted pronunciation resources.

Significance. If the empirical results hold, ATPC would be a useful plug-in for E2E-ASR contextual biasing and a step toward reducing dependence on language-specific pronunciation lexicons. The paper makes a falsifiable claim, includes a head-to-head comparison with a deep-biasing method and a manual-lexicon upper bound, and publicly releases the Aishell-2 ATPC matrix. However, the significance is materially reduced by two issues: the claimed lexicon-free supervision is contradicted by the use of the BABEL pronunciation lexicon to create IPA training targets, and the main quantitative conclusion rests on a single small test set without confidence intervals or significance testing. The cross-corpus setup (ATPC generated on Aishell-2, evaluated on Aishell-1) is a strength if the transfer is shown to be stable, but the paper does not analyze that stability.

major comments (5)
  1. [Abstract, I, III-C] The central claim that ATPC is lexicon-free is contradicted by the experimental setup. Section III-C states: 'Firstly, we convert the transcripts of the BABEL training set into IPA sequences using the pronunciation lexicon of the dataset.' The IPA fine-tuning of XLSR-53 therefore depends on a manually crafted pronunciation lexicon, and Table I shows that the IPA-fine-tuned layer 15 provides the best pronunciation-distinguishing embeddings used for ATPC. Thus the supervision is not merely 'speech and corresponding text annotations' as claimed in the Abstract, and the claim that ATPC 'holds promise for dialects or languages lacking artificial pronunciation lexicons' is overstated. Please either narrow the claim to 'no target-language pronunciation lexicon is required' or provide an experimental variant that does not use any pronunciation lexicon at all.
  2. [III-D] The row-wise normalization description is internally inconsistent. If the ATPC matrix stores DTW distances, the diagonal (self-distance) is the minimum entry in each row, so dividing by the diagonal to set diagonal entries to 1.0 would make off-diagonal entries greater than or equal to 1.0. The text instead says 'any value less than 1.0 signifies a shorter distance' and then selects candidates with 'a distance of less than 1.07.' Please state the exact normalization formula and explain how the 1.07 threshold is applied to the normalized values.
  3. [IV-C, Table II] The empirical claim is based on one test set of 235 sentences with 161 hotwords, with no confidence intervals, significance tests, or multiple runs. The absolute difference between C-g (11.1) and C-g+ATPC (10.3) is 0.8 CER, which may be within decoding noise. Please report bootstrap confidence intervals or paired significance tests for CER, B-CER, and F1, describe the number of decoding runs used, and report sensitivity to the 1.07 threshold on the test set.
  4. [III-D, Table II] The CPPN baseline in Row 3 achieves essentially the same CER as the no-biasing baseline (13.9 vs 13.8), which is unusual for a published deep-biasing method and suggests that its hyperparameters (e.g., deep biasing score 2.0) were not tuned. The claim that ATPC 'consistently outperforms' CPPN is therefore not yet established. Please tune CPPN on the development set or justify the chosen settings by reference to the original paper.
  5. [IV-C] The ATPC matrix is computed on Aishell-2 training speech and applied to the Aishell-1 test set without any analysis of cross-corpus stability of the distance distributions. Since the 1.07 threshold is selected on Aishell-1 development data, it is important to know whether the matrix and threshold transfer to other corpora, speakers, or recording conditions. Please provide evidence such as a comparison of the ATPC matrix computed from Aishell-1 training data, or a threshold sensitivity analysis, to support the generalization claim.
minor comments (6)
  1. [I, IV-C] There are typos: 'vanilla EE-ASR' should be 'vanilla E2E-ASR' and 'Tabel II' should be 'Table II'.
  2. [II-C, Figure 2] Dnorm is used in Figure 2 but never defined; specify whether it is the raw DTW distance or normalized by path length.
  3. [IV-A] The 'Distance Rel Disparity' metric in Table I is not defined in the text; give the formula (presumably (non-homophone distance - homophone distance) / non-homophone distance).
  4. [III-C] The statement that the multilingual speech representation models are trained without manually crafted pronunciation lexicons is misleading in context, since the IPA fine-tuning step uses the BABEL pronunciation lexicon; consider rewording to distinguish pretraining from fine-tuning.
  5. [III-D] Row 0 ('C-g + Manual Lexicons') is not described; explain what the manual-lexicon biasing consists of and how it is combined with the context graph.
  6. [II-C] The random selection of E=100 embeddings should specify the random seed or report variance across seeds for reproducibility.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: ATPC is evaluated against an external benchmark and the threshold is a standard dev-set choice; the lexicon-free claim has an internal consistency issue but not a circular derivation.

full rationale

The derivation chain (ITSE alignment, speech embedding extraction, DTW distance computation, contextual biasing) does not reduce to its own outputs by construction. The ATPC matrix is generated from Aishell-2 speech embeddings and character annotations, while the reported CER/B-CER improvements are measured on the Aishell-1 contextual biasing test set using an independent Wenet model and hotword lists. The 1.07 replacement threshold is selected by sweeping 1.01 to 1.09 on the Aishell-1 development set; this is standard hyperparameter selection, not a fitted parameter renamed as a prediction. The paper cites the authors' prior ITSE work [13], but ITSE is used as an off-the-shelf alignment component and the central claim does not depend on an unverified uniqueness theorem. One consistency concern, not circularity: Section III-C states that the authors 'convert the transcripts of the BABEL training set into IPA sequences using the pronunciation lexicon of the dataset,' which conflicts with the abstract's claim that the supervision is only 'speech and corresponding text annotations' and that ATPC is obtained 'without relying on manually designed pronunciation lexicons.' Since the selected embedding layer (IPA-layer15) is chosen after IPA fine-tuning, the method has an indirect dependency on a manual lexicon. This weakens the lexicon-free framing and is a correctness/claim-support issue, but it does not make the ATPC prediction equivalent to its inputs by construction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the assumption that speech embeddings of the same character are acoustically consistent across contexts and that a distance matrix computed on one Mandarin corpus transfers to another. The generation pipeline itself introduces no new physical entities, but does rely on several unvalidated transfer assumptions.

free parameters (7)
  • ATPC replacement threshold = 1.07
    Tuned on Aishell-1 development set over range 1.01 to 1.09; chosen for lowest CER and second-highest F1. Section III-D.
  • Embedding count per character E = 100
    Design choice; embeddings randomly sampled per character, Section II-C.
  • Minimum occurrence T = 3
    Characters with fewer than 3 occurrences excluded, Section II-C.
  • Embedding layer = layer 15
    Selected based on homophone vs non-homophone disparity in Table I, Section IV-A.
  • Distance function = cosine
    Selected based on Table I; cosine distance gives larger relative disparity, Section IV-A.
  • Context graph score = 6.0
    Hyperparameter for WFST-based baseline, Section III-D.
  • Deep biasing score = 2.0
    Hyperparameter for CPPN baseline, Section III-D.
assumptions (5)
  • domain assumption ITSE provides accurate character-level timestamps
    The entire segmentation of embeddings relies on ITSE timestamps; cited from [13] without re-validation on Aishell-2.
  • domain assumption XLSR-53 embeddings fine-tuned on BABEL IPA capture Mandarin pronunciation distinctions
    The method assumes cross-lingual transfer from BABEL languages to Mandarin; Section III-C.
  • domain assumption DTW distance between same-character embeddings is smaller than between different characters in a way that reflects linguistic pronunciation
    Verified for homophones in Table I, but only as a sanity check, and the transfer to full matrix is assumed.
  • domain assumption The ATPC matrix computed on Aishell-2 training set is applicable to Aishell-1 test set
    Cross-corpus transfer assumed; no analysis of domain mismatch, Section IV-C.
  • domain assumption Pronunciation correlations are stable across utterances, i.e., the average over embeddings represents the character's pronunciation
    Central to the method; the paper does not analyze speaker or context variability.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Automatic Text Pronunciation Correlation Generation and Application for Contextual Biasing." pith.science (2026). https://pith.science/paper/V46YJIJD

@misc{pith2026250100804,
  author       = {Pith},
  title        = {Pith review of: Automatic Text Pronunciation Correlation Generation and Application for Contextual Biasing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V46YJIJD}},
  note         = {Machine review of arXiv:2501.00804}
}
read the original abstract

Effectively distinguishing the pronunciation correlations between different written texts is a significant issue in linguistic acoustics. Traditionally, such pronunciation correlations are obtained through manually designed pronunciation lexicons. In this paper, we propose a data-driven method to automatically acquire these pronunciation correlations, called automatic text pronunciation correlation (ATPC). The supervision required for this method is consistent with the supervision needed for training end-to-end automatic speech recognition (E2E-ASR) systems, i.e., speech and corresponding text annotations. First, the iteratively-trained timestamp estimator (ITSE) algorithm is employed to align the speech with their corresponding annotated text symbols. Then, a speech encoder is used to convert the speech into speech embeddings. Finally, we compare the speech embeddings distances of different text symbols to obtain ATPC. Experimental results on Mandarin show that ATPC enhances E2E-ASR performance in contextual biasing and holds promise for dialects or languages lacking artificial pronunciation lexicons.

Figures

Figures reproduced from arXiv: 2501.00804 by the authors.

Figure 1
Figure 1. The overall diagram of generating ATPC, with c1, c2, and c3 represent multiple embeddings corresponding to the same character in the training [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Pronunciation correlation calculation with DTW. V and W represent [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The visual analysis of generated ATPC matrix subset. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 20 canonical work pages

  1. [1]

    Measuring the gap between hmm- based asr and tts,

    J. Dines, J. Yamagishi, and S. King, “Measuring the gap between hmm- based asr and tts,” IEEE Journal of Selected Topics in Signal Processing , vol. 4, no. 6, pp. 1046–1058, 2010

    work page 2010

  2. [2]

    Thousands of voices for hmm-based speech synthesis–analysis and ap- plication of tts systems built on various asr corpora,

    J. Yamagishi, B. Usabaev, S. King, O. Watts, J. Dines, J. Tian, Y . Guan, R. Hu, K. Oura, Y .-J. Wu, K. Tokuda, R. Karhila, and M. Kurimo, “Thousands of voices for hmm-based speech synthesis–analysis and ap- plication of tts systems built on various asr corpora,” IEEE Transactions on Audio, Speech, and Language Processing , vol. 18, no. 5, pp. 984– 1004, 2010

    work page 2010

  3. [3]

    Keyword search using attention-based end-to-end asr and frame-synchronous phoneme alignments,

    R. Yang, G. Cheng, H. Miao, T. Li, P. Zhang, and Y . Yan, “Keyword search using attention-based end-to-end asr and frame-synchronous phoneme alignments,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 29, pp. 3202–3215, 2021

    work page 2021

  4. [4]

    Modular end-to-end automatic speech recognition framework for acoustic-to- word model,

    Q. Liu, Z. Chen, H. Li, M. Huang, Y . Lu, and K. Yu, “Modular end-to-end automatic speech recognition framework for acoustic-to- word model,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 28, pp. 2174–2183, 2020

    work page 2020

  5. [5]

    Pronunciation dictionary development in resource-scarce environments,

    M. Davel and O. Martirosian, “Pronunciation dictionary development in resource-scarce environments,” 2009

    work page 2009

  6. [6]

    On designing pronunciation lexicons for large vocabulary continuous speech recognition,

    L. Lamel and G. Adda, “On designing pronunciation lexicons for large vocabulary continuous speech recognition,” in Proceeding of F ourth International Conference on Spoken Language Processing. ICSLP ’96 , vol. 1, 1996, pp. 6–9 vol.1

    work page 1996

  7. [7]

    Recent advances in end-to-end automatic speech recognition,

    J. Li, “Recent advances in end-to-end automatic speech recognition,” APSIPA Transactions on Signal and Information Processing , 2021

    work page 2021

  8. [8]

    Eteh: Unified attention-based end-to-end asr and kws architecture,

    G. Cheng, H. Miao, R. Yang, K. Deng, and Y . Yan, “Eteh: Unified attention-based end-to-end asr and kws architecture,” IEEE/ACM Trans- actions on Audio, Speech, and Language Processing , vol. 30, pp. 1360– 1373, 2022

    work page 2022

Show all 26 references
  1. [9]

    Listen, attend and spell: A neural network for large vocabulary conversational speech recognition,

    W. Chan, N. Jaitly, Q. Le, and O. Vinyals, “Listen, attend and spell: A neural network for large vocabulary conversational speech recognition,” in 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2016, pp. 4960–4964

    2016

  2. [10]

    Online hybrid ctc/attention end-to-end automatic speech recognition architecture,

    H. Miao, G. Cheng, P. Zhang, and Y . Yan, “Online hybrid ctc/attention end-to-end automatic speech recognition architecture,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 28, pp. 1452–1465, 2020

    2020

  3. [11]

    Phoebe: Pronunciation-aware contextualization for end-to-end speech recogni- tion,

    A. Bruguier, R. Prabhavalkar, G. Pundak, and T. N. Sainath, “Phoebe: Pronunciation-aware contextualization for end-to-end speech recogni- tion,” in ICASSP 2019-2019 IEEE International Conference on Acous- tics, Speech and Signal Processing (ICASSP) . IEEE, 2019, pp. 6171– 6175

    2019

  4. [12]

    Towards contextual spelling correction for customization of end-to-end speech recognition systems,

    X. Wang, Y . Liu, J. Li, V . Miljanic, S. Zhao, and H. Khalil, “Towards contextual spelling correction for customization of end-to-end speech recognition systems,” IEEE/ACM Trans. Audio, Speech and Lang. Proc. , vol. 30, p. 3089–3097, sep 2022. [Online]. Available: https://doi...

    2022

  5. [13]

    An e2e-asr-based iteratively- trained timestamp estimator,

    R. Yang, G. Cheng, P. Zhang, and Y . Yan, “An e2e-asr-based iteratively- trained timestamp estimator,” IEEE Signal Processing Letters , vol. 29, pp. 1654–1658, 2022

    2022

  6. [14]

    Connection- ist temporal classification: labelling unsegmented sequence data with recurrent neural networks,

    A. Graves, S. Fern ´andez, F. Gomez, and J. Schmidhuber, “Connection- ist temporal classification: labelling unsegmented sequence data with recurrent neural networks,” in Proceedings of the 23rd international conference on Machine learning , 2006, pp. 369–376

    2006

  7. [15]

    Layer-wise analysis of a self-supervised speech representation model,

    A. Pasad, J.-C. Chou, and K. Livescu, “Layer-wise analysis of a self-supervised speech representation model,” in 2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU). IEEE, 2021, pp. 914–921

    2021

  8. [16]

    Dynamic time warping,

    M. M ¨uller, “Dynamic time warping,” Information retrieval for music and motion , pp. 69–84, 2007

    2007

  9. [17]

    Speech recog- nition and keyword spotting for low-resource languages: Babel project research at cued,

    M. J. Gales, K. M. Knill, A. Ragni, and S. P. Rath, “Speech recog- nition and keyword spotting for low-resource languages: Babel project research at cued,” in F ourth International workshop on spoken language technologies for under-resourced languages (SLTU-2014). Internationa...

    2014

  10. [18]

    Aishell-2: Transforming mandarin asr research into industrial scale,

    J. Du, X. Na, X. Liu, and H. Bu, “Aishell-2: Transforming mandarin asr research into industrial scale,” arXiv preprint arXiv:1808.10583 , 2018

    2018 arXiv

  11. [19]

    Aishell-1: An open-source mandarin speech corpus and a speech recognition baseline,

    H. Bu, J. Du, X. Na, B. Wu, and H. Zheng, “Aishell-1: An open-source mandarin speech corpus and a speech recognition baseline,” in 2017 20th conference of the oriental chapter of the international coordinating committee on speech databases and speech I/O systems and assessment...

    2017

  12. [20]

    Contextualized end-to-end speech recognition with contextual phrase prediction network,

    K. Huang, A. Zhang, Z. Yang, P. Guo, B. Mu, T. Xu, and L. Xie, “Contextualized end-to-end speech recognition with contextual phrase prediction network,” in Annual Conference of the International Speech Communication Association, INTERSPEECH 2023 , 2023, pp. 4933– 4937

    2023

  13. [21]

    Un- supervised cross-lingual representation learning for speech recognition,

    A. Conneau, A. Baevski, R. Collobert, A. Mohamed, and M. Auli, “Un- supervised cross-lingual representation learning for speech recognition,” in Proc. Interspeech 2021 , 2021, pp. 2426–2430

    2021

  14. [22]

    Wenet 2.0: More productive end-to-end speech recognition toolkit,

    B. Zhang, D. Wu, Z. Peng, X. Song, Z. Yao, H. Lv, L. Xie, C. Yang, F. Pan, and J. Niu, “Wenet 2.0: More productive end-to-end speech recognition toolkit,” in Proc. Interspeech 2022 , 2022, pp. 1661–1665

    2022

  15. [23]

    Shallow-fusion end-to-end contextual biasing

    D. Zhao, T. N. Sainath, D. Rybach, P. Rondon, D. Bhatia, B. Li, and R. Pang, “Shallow-fusion end-to-end contextual biasing.” in Proc. Interspeech 2019 , 2019, pp. 1418–1422

    2019

  16. [24]

    Improved neural language model fusion for stream- ing recurrent neural network transducer,

    S. Kim, Y . Shangguan, J. Mahadeokar, A. Bruguier, C. Fuegen, M. L. Seltzer, and D. Le, “Improved neural language model fusion for stream- ing recurrent neural network transducer,” in ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (IC...

    2021

  17. [25]

    Cif-based collaborative decoding for end-to-end contextual speech recognition,

    M. Han, L. Dong, S. Zhou, and B. Xu, “Cif-based collaborative decoding for end-to-end contextual speech recognition,” in ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2021, pp. 6528–6532

    2021

  18. [26]

    Joint grapheme and phoneme embeddings for contextual end-to-end asr

    Z. Chen, M. Jain, Y . Wang, M. L. Seltzer, and C. Fuegen, “Joint grapheme and phoneme embeddings for contextual end-to-end asr.” in Proc. Interspeech 2019 , 2019, pp. 3490–3494

    2019

Pith tools

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