Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

ASR-EC Benchmark: Evaluating Large Language Models on Chinese ASR Error Correction

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that on the first Chinese ASR error correction benchmark, ASR-EC, a multimodal LLM that hears the audio alongside the transcript achieves the lowest character error rates, beating prompting, text-only finetuning, and the…

desk verdict A plausible first Chinese ASR error-correction benchmark, undermined by a central Table 4 inconsistency that makes the multimodal SOTA claim unverifiable as written. read the letter →

arxiv 2412.03075 v1 pith:DLYV56I4 submitted 2024-12-04 cs.CL cs.SDeess.AS

Victor Junqiu Wei , Weicheng Wang , Di Jiang , Yuanfeng Song , Lu Wang This is my paper · ORCID
classification cs.CLcs.SDeess.AS
keywords ASRerrorcorrectionChinesespeechrecognitionlargelanguagemodelsmultimodalaudio-textmodelbenchmarkdatasetcharacterrateLoRAfinetuning
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

The paper sets out to give Chinese ASR error correction a public testbed and to find out how large language models should be used on it. It builds ASR-EC by decoding audio from four Chinese speech corpora with two different ASR pipelines, yielding paired erroneous and correct transcripts across short and long utterances, and claims this is the first Chinese ASR error correction benchmark. Using that benchmark, it compares three LLM strategies: prompting, LoRA finetuning, and multimodal augmentation that feeds the audio together with the transcript into the model. The headline result is that prompting alone raises the error rate, finetuning helps some models, and the multimodal approach achieves the lowest character error rates, beating both the ASR baselines and text-only correction. If the benchmark is representative, the practical route for LLM-based ASR post-processing in Chinese is to give the model access to the audio signal.

What carries the argument

The machine that carries the argument is ASR-EC, a paired dataset of audio clips, erroneous transcripts from two ASR pipelines, and correct transcripts, split into short and long utterances and annotated by substitution, deletion, and insertion errors. The method that produces the headline result is multimodal augmentation: the audio is encoded and concatenated with an error-correction instruction and the erroneous text, and a multimodal LLM is finetuned end-to-end to output the corrected text. The audio channel lets the model resolve homophone and near-homophone confusions that are invisible from text alone, which is why the paper finds audio-plus-text correction superior to both prompting and text-only finetuning.

What would settle it

Take a third Chinese ASR system not used in the benchmark, run its transcripts through the same multimodal finetuning protocol, and compare character error rates against text-only finetuning; if the audio-based gains disappear or reverse, the central claim is an artifact of the two chosen ASR engines.

Watch

Extended reading notes

Core claim

The paper's central claim is that a multimodal LLM finetuned end-to-end on ASR-EC—presented with the raw audio and the erroneous transcript jointly—corrects Chinese ASR errors better than any prompting or text-only finetuning configuration tested. The authors report that prompting zero-shot, few-shot, or multi-step consistently increases character error rate relative to the original ASR output because LLMs over-correct. LoRA finetuning of text-only LLMs recovers most of the gap but leaves the models short of the multimodal result, with Baichuan2 the strongest of the three text models. Finetuning the audio-language model on ASR-EC also improves on its out-of-the-box ASR mode, indicating the benchmark provides a useful supervised signal even for models already pretrained on audio. The authors also identify a residual floor: errors involving names and pronouns cannot be fixed without external context or prior knowledge.

Load-bearing premise

The benchmark's error mix is assumed representative because it comes from two specific ASR engines and keeps only 10% of error-free transcripts; if that mix is unrepresentative, the relative gains of the multimodal method may not generalize.

Editorial extensions

If this is right

  • Prompting a general LLM to correct ASR output without training data is counterproductive: it raises character error rate on both benchmark sets.
  • Parameter-efficient finetuning with LoRA turns text-only LLMs into usable correctors, but which models benefit varies and none reach the multimodal result.
  • Giving a multimodal LLM the audio alongside the transcript yields the largest character error rate reductions across short, long, and mixed utterances.
  • Finetuning an audio-language model on ASR-EC improves its correction performance over its default ASR mode, so the benchmark is useful even for audio-pretrained models.
  • Name and pronoun errors remain a floor that neither text nor audio context in the benchmark can fix.

Reading between the lines

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

  • Beyond the paper: because the multimodal gains come from acoustic disambiguation, production Chinese ASR pipelines could treat a finetuned audio-language model as a post-processing stage whenever the original audio is still available.
  • Beyond the paper: the reported name and pronoun floor suggests that coupling correction with external knowledge sources, such as contact lists or entity databases, is a concrete next test that the paper does not run.
  • Beyond the paper: the 10% retention cap on error-free utterances means the benchmark over-represents erroneous speech; a deployment where most transcripts are already correct would need different precision and regression metrics to judge correction value.
  • Beyond the paper: since only two ASR engines generated the errors, extending the benchmark with transcripts from additional architectures would test whether the multimodal advantage holds across error distributions.
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 introduces ASR-EC, a Chinese ASR error correction benchmark constructed by decoding audio from THCHS-30, AISHELL-1, AISHELL-2, and WeNetSpeech with two Kaldi-based pipelines (Kaldi-K1 and Kaldi-K2), splitting the resulting transcripts into short and long utterance subsets, and retaining only 10% of error-free utterances. It then evaluates three paradigms for applying LLMs to ASR error correction: prompting (zero-shot, three-shot, and multi-step), LoRA fine-tuning, and a multimodal augmentation approach that concatenates encoded raw audio with the instruction and the erroneous ASR text as input to the LLM. The reported findings are that prompting degrades CER relative to the ASR baselines, LoRA fine-tuning improves a subset of models, and multimodal augmentation achieves the lowest CERs, with an additional experiment showing that fine-tuning Qwen-Audio on ASR-EC improves over Qwen-Audio's ASR mode. The paper claims to present the first Chinese ASR error correction benchmark and the first study of LLM-based ASR error correction for Chinese.

Significance. If the benchmark construction is sound and the multimodal results are reproducible, ASR-EC would be a useful resource for the community: it aggregates four widely used Chinese speech corpora, provides error breakdowns by substitution/deletion/insertion, and covers both hybrid and end-to-end ASR decoding pipelines. The paper also provides a systematic comparison of prompting, fine-tuning, and multimodal input strategies, which is valuable for practitioners choosing among these paradigms. However, the central comparative claim that multimodal augmentation is the most effective approach is not currently supported because the multimodal rows in Table 4 are not tied to a concrete audio-capable model for Baichuan2, ChatGLM3, and Qwen, and the Qwen-Audio experiment in Section 7.4 does not isolate the contribution of the audio modality. The benchmark's representativeness is also asserted rather than demonstrated, so the generality of the conclusions remains uncertain.

major comments (5)
  1. [§7.2, Table 4 and §6] The central claim that 'multi-modal augmentation ... achieves state-of-the-art performance' rests on rows in Table 4 that are internally inconsistent with the model descriptions in Table 3 and Section 6. Baichuan2, ChatGLM3, and Qwen are listed as text-only models in Table 3, yet Table 4 reports 'Multimodal' CER values for these models (e.g., 5.96 for Baichuan2 on ASR-EC A* mixed utterances). Section 6 defines the multimodal input as encoded raw audio concatenated with the instruction and erroneous ASR text, but no audio encoder, adapter, or input projection is described for these three checkpoints. The authors must either specify the audio-capable models or pipeline used to produce those rows, including any additional parameters or training steps, or remove/relabel them. As written, the comparative conclusion that multimodal augmentation is the most effective paradigm is unsupported.
  2. [§7.4, Table 5] The Qwen-Audio experiment does not test whether adding audio is the decisive factor. Table 5 compares Qwen-Audio in its ASR mode with Qwen-Audio fine-tuned on ASR-EC; both variants use audio and text inputs. To support the claim that audio plus transcript is more effective than transcript alone, the paper needs a controlled comparison with a text-only LoRA fine-tuned model (e.g., Qwen or Qwen-Audio with audio inputs ablated) trained on the same ASR-EC data. Without such a comparison, the observed improvement may simply reflect fine-tuning on the training distribution rather than the multimodal input format.
  3. [§3.2] The benchmark construction uses an arbitrary 10% retention rate for utterances with CER 0 ('To control the number of those with a CER of 0, only 10% of them are kept'), but no sensitivity analysis or justification is given for this rate. Changing this rate can substantially alter the error distribution and the difficulty of the correction task, which directly affects all downstream comparisons. Additionally, the claim that Kaldi-K1 and Kaldi-K2 are 'the only two pipelines in the history of ASR systems' is unsupported and implausible, and the assertion that their error patterns are representative of industry-grade Chinese ASR systems is not backed by evidence. The authors should provide a justification for the 10% cap and temper or support the representativeness claim.
  4. [Tables 4 and 5] All CER values are reported as point estimates without confidence intervals, significance tests, or multiple-run variability. Several comparisons that drive the conclusions involve small absolute differences (e.g., Baichuan2 vs Qwen multimodal on ASR-EC A* long utterances: 6.99 vs 5.64), and the ranking of models and paradigms may change within noise. The authors should report error bars or paired significance tests, or at least state the number of evaluation runs and the variance observed.
  5. [Abstract, Section 2, Section 3.1] The paper claims to present 'the first Chinese ASR error correction benchmark' and 'the first one to study LLM-based ASR error correction,' but Section 2 itself cites earlier LLM-based ASR error correction work (references [10]–[15]) and the related-work discussion indicates substantial prior art. The novelty claims need to be qualified and reconciled with these references; as stated, they overstate the contribution.
minor comments (6)
  1. [Table 2] There is a typo in the Substitution CER for ASR-EC B* short utterances: '9,59' should be '9.59'. In addition, the relationship between the test-set CER in the upper part of the table and the component CERs in the lower part is not explained (e.g., ASR-EC A* Whole Dataset has Overall CER 13.32 while the reported test CER is 12.42); the authors should clarify which split the component CERs are computed on.
  2. [Section 4] The sentence 'Prompting is an emerging technique for fine-tuning large language models' is inaccurate in standard terminology: prompting does not modify model parameters. Please rephrase to describe prompting as an inference-time adaptation method.
  3. [Section 7.2] The text refers to a 'one-shot' setting, but Table 4 reports 'Three-Shot.' Please make the terminology consistent and describe the exact number and selection of few-shot examples.
  4. [Section 7.1 and References] Reference [2] is cited as 'Kaldi-K1' but the actual reference is the Kaldi toolkit paper; the paper should cite or specify the specific Kaldi-K1 recipe (e.g., the multistream CNN configuration) and the Kaldi-K2/Zipformer-Transducer source separately.
  5. [Contributions and Section 5] There is a typo in the contributions list: 'funetuning' should be 'finetuning.' Also, the terms 'fine-tuning' and 'funetuning' are used inconsistently throughout the paper.
  6. [Reference [1]] Reference [1] points to an anonymous GitHub repository path containing 'train_data' only, while the text states the dataset will be released upon publication. The availability status of the full benchmark (including test sets and evaluation scripts) should be clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the benchmark construction and all CER comparisons are empirical measurements, not derivations that reduce to their inputs.

full rationale

The paper is an empirical benchmark and evaluation study rather than a derivation. The ASR-EC dataset is constructed from external speech corpora (THCHS-30, AISHELL-1, AISHELL-2, WeNetSpeech) and two Kaldi-based ASR pipelines; the resulting CER values are measured, not optimized to match a target conclusion. The three LLM paradigms are evaluated on held-out test splits, and no fitted parameter is later renamed as a prediction. The 'first Chinese ASR error correction benchmark' statement is a novelty claim, not a logical loop. The reviewer-flagged inconsistency that Table 4 reports 'Multimodal' results for models that Table 3 labels as text-only is a serious reproducibility and internal-consistency concern, but it is not circularity: those rows, if valid, would come from an undeclared experimental setup rather than from a derivation that is equivalent to its own input. Similarly, the decision to keep only 10% of CER=0 utterances is a benchmark design choice, not a parameter that by construction forces the later comparative conclusions. Self-citations appear only in related-work and methodology references and are not load-bearing for the central empirical claim. Therefore no circular step meeting the stated evidentiary standard is present; score 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No invented entities are introduced. The empirical results rest on the benchmark's composition (two Kaldi ASR pipelines, 10% cap on error-free utterances, manual transcripts as ground truth, 1,024-utterance test subsets) and on the assumption that the multimodal integration described in Figure 2 was correctly implemented for the models in Table 4. These are domain assumptions rather than fitted quantities.

free parameters (4)
  • CER=0 retention rate = 10%
    Section 3.2: 'only 10% of them are kept' for utterances with CER=0; chosen by hand, alters the difficulty mix of the benchmark.
  • Few-shot example count = 3
    Section 4: three input-output pairs are used in few-shot prompting; not justified.
  • ChatGLM3 repetition penalty = 1.05
    Section 7.1: adjusted from default 1.0 to avoid repetitive generation; affects all ChatGLM3 results.
  • LoRA fine-tuning hyperparameters = not specified
    Section 7.1: hyperparameters are taken from official and third-party repositories but not listed, so the finetuning results are not reproducible from the paper alone.
assumptions (4)
  • domain assumption Manual transcripts in THCHS-30, AISHELL-1, AISHELL-2, and WeNetSpeech are accurate ground truth.
    The CER evaluation in all tables compares model outputs to these transcripts (Section 3.1); any transcript errors propagate into every reported number.
  • domain assumption Kaldi-K1 and Kaldi-K2 are representative of industry-grade Chinese ASR systems.
    ASR-EC is built exclusively from these two pipelines (Section 3.2); the paper claims the benchmark captures a 'wide spectrum' of ASR errors, but no external validation against other ASR systems is provided.
  • domain assumption The 1,024-utterance test subsets are representative of the benchmark's distribution.
    All reported CERs come from these small test sets (Table 2), yet the sampling procedure and confidence intervals are not given.
  • domain assumption The multimodal experiments in Table 4 correctly integrate audio into the listed text-only LLMs.
    The paper does not specify how Baichuan2, ChatGLM3, and Qwen receive audio input; the architecture in Figure 2 is described generically (Section 6).

how reviews work

0 comments
Cite this review

Pith. "Pith review of ASR-EC Benchmark: Evaluating Large Language Models on Chinese ASR Error Correction." pith.science (2026). https://pith.science/paper/DLYV56I4

@misc{pith2026241203075,
  author       = {Pith},
  title        = {Pith review of: ASR-EC Benchmark: Evaluating Large Language Models on Chinese ASR Error Correction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DLYV56I4}},
  note         = {Machine review of arXiv:2412.03075}
}
read the original abstract

Automatic speech Recognition (ASR) is a fundamental and important task in the field of speech and natural language processing. It is an inherent building block in many applications such as voice assistant, speech translation, etc. Despite the advancement of ASR technologies in recent years, it is still inevitable for modern ASR systems to have a substantial number of erroneous recognition due to environmental noise, ambiguity, etc. Therefore, the error correction in ASR is crucial. Motivated by this, this paper studies ASR error correction in the Chinese language, which is one of the most popular languages and enjoys a large number of users in the world. We first create a benchmark dataset named \emph{ASR-EC} that contains a wide spectrum of ASR errors generated by industry-grade ASR systems. To the best of our knowledge, it is the first Chinese ASR error correction benchmark. Then, inspired by the recent advances in \emph{large language models (LLMs)}, we investigate how to harness the power of LLMs to correct ASR errors. We apply LLMs to ASR error correction in three paradigms. The first paradigm is prompting, which is further categorized as zero-shot, few-shot, and multi-step. The second paradigm is finetuning, which finetunes LLMs with ASR error correction data. The third paradigm is multi-modal augmentation, which collectively utilizes the audio and ASR transcripts for error correction. Extensive experiments reveal that prompting is not effective for ASR error correction. Finetuning is effective only for a portion of LLMs. Multi-modal augmentation is the most effective method for error correction and achieves state-of-the-art performance.

Figures

Figures reproduced from arXiv: 2412.03075 by the authors.

Figure 1
Figure 1. Pipelines for Erroneous ASR Transcripts Construction [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. ASR error correction with Multimodal model. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Outputs of Prompting Approaches By fine-tuning the open-source large models, the performance of all fine-tuned LLMs showed significant improvement, as revealed in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Outputs of different fine-tuned models [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: LLMs are not able to correct some types of errors without context or prior knowledge. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

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

  1. Contextualized Token Discrimination for Speech Search Query Correction

    cs.SD 2025-09 reject novelty 4.0 of 10

    CTD uses BERT token representations plus a composition layer to correct Chinese spelling errors in ASR queries, but the reported gains lack matched baselines and released data.

Reference graph

Works this paper leans on

53 extracted references · 31 canonical work pages · cited by 1 Pith paper

  1. [10]

    N-best t5: Robust asr error correction using multiple input hypotheses and constrained decoding space,

    R. Ma, M. J. Gales, K. M. Knill, and M. Qian, “N-best t5: Robust asr error correction using multiple input hypotheses and constrained decoding space,” arXiv preprint arXiv:2303.00456, 2023

    arXiv 2023

  2. [15]

    It’s never too late: Fusing acoustic information into large language models for automatic speech recognition,

    C. Chen, R. Li, Y . Hu, S. M. Siniscalchi, P.-Y . Chen, E. Chng, and C.-H. H. Yang, “It’s never too late: Fusing acoustic information into large language models for automatic speech recognition,” 2024

    work page 2024

  3. [1]

    [Online]

    Data, 2024. [Online]. Available: https://github.com/anonymity139/ASR-EC-Benchmark/tree/main/ASR/train_data

    work page 2024

  4. [2]

    The kaldi speech recognition toolkit,

    D. Povey, A. Ghoshal, G. Boulianne, L. Burget, O. Glembek, N. Goel, M. Hannemann, P. Motlicek, Y . Qian, P. Schwarz, J. Silovsky, G. Stemmer, and K. Vesely, “The kaldi speech recognition toolkit,” in IEEE 2011 Workshop on Automatic Speech Recognition and Understanding. IEEE Signal Processing Society, Dec. 2011, iEEE Catalog No.: CFP11SRW-USB

    work page 2011

  5. [3]

    BERT: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” in Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), J. Burstein, C. Doran, and T. Solorio, Eds...

    work page 2019

  6. [4]

    Language models are few-shot learners,

    T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell et al., “Language models are few-shot learners,” Advances in neural information processing systems, vol. 33, pp. 1877–1901, 2020

    1901

  7. [5]

    A survey of large language models,

    W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Donget al., “A survey of large language models,” arXiv preprint arXiv:2303.18223, 2023

    arXiv 2023

  8. [6]

    Lora: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “Lora: Low-rank adaptation of large language models,” arXiv preprint arXiv:2106.09685, 2021

    arXiv 2021

Show all 53 references
  1. [7]

    Correction of automatic speech recognition with transformer sequence- to-sequence model,

    O. Hrinchuk, M. Popova, and B. Ginsburg, “Correction of automatic speech recognition with transformer sequence- to-sequence model,” in Icassp 2020-2020 ieee international conference on acoustics, speech and signal processing (icassp). IEEE, 2020, pp. 7074–7078

    2020

  2. [8]

    Bart based semantic correction for mandarin automatic speech recognition system,

    Y . Zhao, X. Yang, J. Wang, Y . Gao, C. Yan, and Y . Zhou, “Bart based semantic correction for mandarin automatic speech recognition system,” arXiv preprint arXiv:2104.05507, 2021

    2021 arXiv

  3. [9]

    Jiang, C

    D. Jiang, C. Zhang, and Y . Song,Probabilistic Topic Models: Foundation and Application. Springer, 2023

    2023

  4. [11]

    Generative speech recognition error correction with large language models and task-activating prompting,

    C.-H. H. Yang, Y . Gu, Y .-C. Liu, S. Ghosh, I. Bulyko, and A. Stolcke, “Generative speech recognition error correction with large language models and task-activating prompting,” in 2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) . IEEE, Dec. 2023. [On...

    2023

  5. [12]

    Can generative large language models perform asr error correction?

    R. Ma, M. Qian, P. Manakul, M. Gales, and K. Knill, “Can generative large language models perform asr error correction?” arXiv preprint arXiv:2307.04172, 2023. 10 A PREPRINT - DECEMBER 5, 2024

    2023 arXiv

  6. [13]

    Large language models are efficient learners of noise-robust speech recognition,

    Y . Hu, C. Chen, C.-H. H. Yang, R. Li, C. Zhang, P.-Y . Chen, and E. Chng, “Large language models are efficient learners of noise-robust speech recognition,” 2024

    2024

  7. [14]

    Qwen-audio: Advancing universal audio understanding via unified large-scale audio-language models,

    Y . Chu, J. Xu, X. Zhou, Q. Yang, S. Zhang, Z. Yan, C. Zhou, and J. Zhou, “Qwen-audio: Advancing universal audio understanding via unified large-scale audio-language models,” 2023

    2023

  8. [16]

    Predicting the next app that you are going to use,

    R. Baeza-Yates, D. Jiang, F. Silvestri, and B. Harrison, “Predicting the next app that you are going to use,” in Proceedings of the eighth ACM international conference on web search and data mining, 2015, pp. 285–294

    2015

  9. [17]

    A neural probabilistic language model,

    Y . Bengio, R. Ducharme, P. Vincent, and C. Jauvin, “A neural probabilistic language model,”Journal of machine learning research, vol. 3, no. Feb, pp. 1137–1155, 2003

    2003

  10. [18]

    Efficient estimation of word representations in vector space,

    T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” arXiv preprint arXiv:1301.3781, 2013

    2013 arXiv

  11. [19]

    Glove: Global vectors for word representation,

    J. Pennington, R. Socher, and C. D. Manning, “Glove: Global vectors for word representation,” in Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 2014, pp. 1532–1543

    2014

  12. [20]

    Densely connected cnn with multi-scale feature attention for text classification

    S. Wang, M. Huang, and Z. Deng, “Densely connected cnn with multi-scale feature attention for text classification.” in IJCAI, 2018, pp. 4468–4474

    2018

  13. [21]

    Pretraining enhanced rnn transducer,

    J. Lu, R. Lian, D. Jiang, Y . Song, Z. Su, V . J. Wei, and L. Yang, “Pretraining enhanced rnn transducer,”CAAI Artificial Intelligence Research, vol. 3, 2024

    2024

  14. [22]

    A phonetic-semantic pre-training model for robust speech recognition,

    X. Wu, R. Lian, D. Jiang, Y . Song, W. Zhao, Q. Xu, and Q. Yang, “A phonetic-semantic pre-training model for robust speech recognition,” CAAI Artificial Intelligence Research, vol. 1, no. 1, 2022

    2022

  15. [23]

    Multi-turn response selection for chatbots with deep attention matching network,

    X. Zhou, L. Li, D. Dong, Y . Liu, Y . Chen, W. X. Zhao, D. Yu, and H. Wu, “Multi-turn response selection for chatbots with deep attention matching network,” in Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2018...

    2018

  16. [24]

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

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

    2018 arXiv

  17. [25]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, and I. Sutskever, “Language models are unsupervised multitask learners,” OpenAI blog, vol. 1, no. 8, p. 9, 2019

    2019

  18. [26]

    Albert: A lite bert for self-supervised learning of language representations,

    Z. Lan, M. Chen, S. Goodman, K. Gimpel, P. Sharma, and R. Soricut, “Albert: A lite bert for self-supervised learning of language representations,” arXiv preprint arXiv:1909.11942, 2019

    1909 arXiv

  19. [27]

    Roberta: A robustly optimized bert pretraining approach,

    Y . Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V . Stoyanov, “Roberta: A robustly optimized bert pretraining approach,” arXiv preprint arXiv:1907.11692, 2019

    1907 arXiv

  20. [28]

    Training multilingual pre-trained language model with byte-level subwords,

    J. Wei, Q. Liu, Y . Guo, and X. Jiang, “Training multilingual pre-trained language model with byte-level subwords,” arXiv preprint arXiv:2101.09469, 2021

    2021 arXiv

  21. [29]

    Nezha: Neural contextualized representation for chinese language understanding,

    J. Wei, X. Ren, X. Li, W. Huang, Y . Liao, Y . Wang, J. Lin, X. Jiang, X. Chen, and Q. Liu, “Nezha: Neural contextualized representation for chinese language understanding,” arXiv preprint arXiv:1909.00204, 2019

    1909 arXiv

  22. [30]

    Heterogeneous latent topic discovery for semantic text mining,

    Y . Li, D. Jiang, R. Lian, X. Wu, C. Tan, Y . Xu, and Z. Su, “Heterogeneous latent topic discovery for semantic text mining,” IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 1, pp. 533–544, 2021

    2021

  23. [31]

    Industrial federated topic modeling,

    D. Jiang, Y . Tong, Y . Song, X. Wu, W. Zhao, J. Peng, R. Lian, Q. Xu, and Q. Yang, “Industrial federated topic modeling,” ACM Transactions on Intelligent Systems and Technology (TIST), vol. 12, no. 1, pp. 1–22, 2021

    2021

  24. [32]

    Memetic federated learning for biomedical natural language processing,

    X. Zhou, C. Tan, D. Jiang, B. Zhang, S. Li, Y . Xu, Q. Xu, and S. Gao, “Memetic federated learning for biomedical natural language processing,” in Natural Language Processing and Chinese Computing: 10th CCF International Conference, NLPCC 2021, Qingdao, China, October 13–17, 2...

    2021

  25. [33]

    Federated topic modeling,

    D. Jiang, Y . Song, Y . Tong, X. Wu, W. Zhao, Q. Xu, and Q. Yang, “Federated topic modeling,” inProceedings of the 28th ACM international conference on information and knowledge management, 2019, pp. 1071–1080

    2019

  26. [34]

    Topicocean: An ever-increasing topic model with meta-learning,

    Y . Song, Y . Tong, S. Bao, D. Jiang, H. Wu, and R. C.-W. Wong, “Topicocean: An ever-increasing topic model with meta-learning,” in 2020 IEEE International Conference on Data Mining (ICDM). IEEE, 2020, pp. 1262–1267

    2020

  27. [35]

    Expanding chatbot knowledge in customer service: Context-aware similar question generation using large language models,

    M. Hong, Y . Song, D. Jiang, L. Wang, Z. Guo, and C. J. Zhang, “Expanding chatbot knowledge in customer service: Context-aware similar question generation using large language models,” arXiv preprint arXiv:2410.12444, 2024

    2024

  28. [36]

    Latent topic embedding,

    D. Jiang, L. Shi, R. Lian, and H. Wu, “Latent topic embedding,” in Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, 2016, pp. 2689–2698. 11 A PREPRINT - DECEMBER 5, 2024

    2016

  29. [37]

    Deep contextualized word representations,

    M. E. Peters, M. Neumann, M. Iyyer, M. Gardner, C. Clark, K. Lee, and L. Zettlemoyer, “Deep contextualized word representations,” arXiv preprint arXiv:1802.05365, 2018

    2018 arXiv

  30. [38]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, “Attention is all you need,” arXiv preprint arXiv:1706.03762, 2017

    2017 arXiv

  31. [39]

    Continuous self-attention models with neural ode networks,

    J. Zhang, P. Zhang, B. Kong, J. Wei, and X. Jiang, “Continuous self-attention models with neural ode networks,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 16, 2021, pp. 14 393–14 401

    2021

  32. [40]

    Tensorcoder: Dimension-wise attention via tensor representation for natural language modeling,

    S. Zhang, P. Zhang, X. Ma, J. Wei, N. Wang, and Q. Liu, “Tensorcoder: Dimension-wise attention via tensor representation for natural language modeling,” arXiv preprint arXiv:2008.01547, 2020

    2008 arXiv

  33. [41]

    Clusterformer: Neural clustering attention for efficient and effective transformer,

    N. Wang, G. Gan, P. Zhang, S. Zhang, J. Wei, Q. Liu, and X. Jiang, “Clusterformer: Neural clustering attention for efficient and effective transformer,” in Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2022, pp...

    2022

  34. [42]

    Hypoformer: Hybrid decomposition transformer for edge-friendly neural machine translation,

    S. Li, P. Zhang, G. Gan, X. Lv, B. Wang, J. Wei, and X. Jiang, “Hypoformer: Hybrid decomposition transformer for edge-friendly neural machine translation,” in Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, 2022, pp. 7056–7068

    2022

  35. [43]

    Improving language understanding by generative pre-training,

    A. Radford, K. Narasimhan, T. Salimans, and I. Sutskever, “Improving language understanding by generative pre-training,” 2018

    2018

  36. [44]

    Thchs-30: A free chinese speech corpus,

    D. Wang and X. Zhang, “Thchs-30: A free chinese speech corpus,” arXiv preprint arXiv:1512.01882, 2015

    2015 arXiv

  37. [45]

    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

  38. [46]

    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

  39. [47]

    Wenetspeech: A 10000+ hours multi-domain mandarin corpus for speech recognition,

    B. Zhang, H. Lv, P. Guo, Q. Shao, C. Yang, L. Xie, X. Xu, H. Bu, X. Chen, C. Zeng et al., “Wenetspeech: A 10000+ hours multi-domain mandarin corpus for speech recognition,” in ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). I...

    2022

  40. [48]

    Glm: General language model pretraining with autoregressive blank infilling,

    Z. Du, Y . Qian, X. Liu, M. Ding, J. Qiu, Z. Yang, and J. Tang, “Glm: General language model pretraining with autoregressive blank infilling,” inProceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2022, pp. 320–335

    2022

  41. [49]

    Qwen technical report,

    J. Bai, S. Bai, Y . Chu, Z. Cui, K. Dang, X. Deng, Y . Fan, W. Ge, Y . Han, F. Huanget al., “Qwen technical report,” arXiv preprint arXiv:2309.16609, 2023

    2023 arXiv

  42. [50]

    Baichuan 2: Open large-scale language models,

    A. Yang, B. Xiao, B. Wang, B. Zhang, C. Bian, C. Yin, C. Lv, D. Pan, D. Wang, D. Yanet al., “Baichuan 2: Open large-scale language models,” arXiv preprint arXiv:2309.10305, 2023

    2023 arXiv

  43. [51]

    DB-GPT-Hub,

    D.-G.-H. Team, “DB-GPT-Hub,” 2023

    2023

  44. [52]

    Llama factory,

    hiyouga, “Llama factory,” https://github.com/hiyouga/LLaMA-Factory, 2023

    2023

  45. [53]

    Chinese spelling correction as rephrasing language model,

    L. Liu, H. Wu, and H. Zhao, “Chinese spelling correction as rephrasing language model,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 17, 2024, pp. 18 662–18 670. 12

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.