Pith. sign in

REVIEW 4 major objections 8 minor 2 cited by

Survey of End-to-End Multi-Speaker Automatic Speech Recognition for Monaural Audio

T0 review · 4 major / 8 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This survey argues that after six years of end-to-end multi-speaker ASR research, no single architecture consistently wins, and real-world benchmark progress on the AMI meeting corpus has not improved since 2021.

desk verdict Useful survey with a solid SIMO/SISO taxonomy, but the AMI stagnation claim rests on a table that compares pre-segmented and long-form systems as if they were the same protocol. read the letter →

arxiv 2505.10975 v3 pith:ZDGYNS6Q submitted 2025-05-16 cs.CL cs.AIcs.SDeess.AS

Xinlu He , Jacob Whitehill This is my paper · ORCID
classification cs.CLcs.AIcs.SDeess.AS
keywords multi-speakerASRend-to-endserializedoutputtrainingSIMOSISOspeakerdiarizationlong-formspeechrecognitioncpWER
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 survey asks whether end-to-end neural systems that map a single microphone's mixed audio directly to speaker-attributed transcripts have converged on a winning design. The paper organizes the field into two architectural families—SIMO, which splits the mixture into one output branch per speaker, and SISO, which writes all speakers into one serialized transcript—and compares recent systems on standard benchmarks. Its sharpest finding is negative: no architecture family consistently beats the other, and on the real-world AMI meeting corpus the best reported score remains a 2021 model trained on 900k hours of simulated mixtures. The survey argues the field is shifting from chasing benchmark word-error rates toward scenario-specific modeling, foundation-model adaptation, and long-form transcription. A reader should care because the paper turns a scattered literature into a map and identifies where progress has, or has not, actually occurred.

What carries the argument

The organizing device is the SIMO/SISO dichotomy. In single-input multiple-output (SIMO) designs, a shared encoder processes the mixed audio and parallel branches each decode one speaker's transcript, trained with permutation-invariant losses to resolve output order. In single-input single-output (SISO) designs, serialized output training collapses all speakers into one token stream, with special tokens marking speaker changes; this relaxes the fixed-speaker-count constraint and enables direct fine-tuning of single-speaker foundation models. The comparison machinery is the evaluation table, which reports cpWER (concatenated minimum-permutation WER) alongside stated metric caveats, and the paper's taxonomy of improvements—separation enhancement, dynamic speaker handling, pretraining, multi-task learning, and multimodal conditioning—that structures the review.

What would settle it

Evaluate the 2021 50M-parameter pretrained model and a recent strong system (for example, one of the Whisper-adapted SIMO systems) on the same AMI SDM eval split, computing cpWER with identical segmentation and scoring. If the recent system's cpWER falls clearly below the 2021 model's 21.2 with a matched test, the stagnation claim would be refuted; if the scores tie within metric noise, the claim would be supported.

Watch

Extended reading notes

Core claim

The paper's central claim is that after six years of end-to-end multi-speaker ASR research, no single architecture—SIMO or SISO—consistently outperforms the others, and the real-world benchmark numbers show no steady improvement since 2021. On AMI, the best published cpWER was achieved by a small (50M parameter) model pretrained on 900k hours of simulated mixture data, reported in 2021; recent proposals adapt large foundation models, add auxiliary losses, or inject speaker conditioning, yet they do not uniformly surpass that result. The authors therefore characterize contemporary work as less concerned with marginal WER gains and more with scenario-dependent factors, novel information fusion, and lightweight adaptation of pretrained single-speaker systems. They also document that fair comparison is hindered by inconsistent metrics (WER vs cpWER vs SA-WER), microphone settings, and split choices, which they report side by side rather than averaging.

Load-bearing premise

The survey's cross-method conclusions rest on the assumption that the numbers collected from different papers—using different metrics, microphone settings, and development/evaluation splits—can be meaningfully compared as if they were produced under one protocol.

Editorial extensions

If this is right

  • If the stagnation observation holds, benchmark-driven claims of progress in multi-speaker ASR must be reexamined: gains reported on one dataset or with one metric may not transfer to real meeting recordings.
  • SIMO and SISO should be treated as design choices with different trade-offs (fixed versus variable speaker count, modularity versus joint modeling), not as a progression where one supersedes the other.
  • Foundation-model adaptation is the most promising shared route: both frameworks report strong results while training only about 8–10% of parameters.
  • For long-form audio, the bottleneck moves from recognition to segmentation and speaker-consistent hypothesis stitching, so improvements there would directly raise end-to-end system usability.
  • Comparison quality depends on community-wide standardization; the paper's own table shows why mixed metrics and splits prevent reliable cross-method ranking.

Reading between the lines

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

  • Because the AMI leader is a 2021 model trained on vast simulated data, the survey's own evidence suggests data scale and simulation realism, not architecture, may be the binding constraint; a controlled study that holds architecture fixed and scales simulated training hours would test this directly.
  • The taxonomy's edge cases—the same system labeled 'Hybrid' on one panel and 'SIMO' on another—imply that a unified labeling scheme would need to classify by mechanism (explicit separation versus serialization) rather than by model family name.
  • The metric mixing documented in the survey could be resolved by a shared evaluation harness that recomputes cpWER on common splits; until then, any single-number reading of the table (including the stagnation claim) is provisional.
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

4 major / 8 minor

Summary. This paper is a survey of end-to-end multi-speaker automatic speech recognition for monaural audio. It organizes recent work into a taxonomy of single-input multiple-output (SIMO) and single-input single-output (SISO) architectures, reviews improvements in separation, speaker conditioning, multi-task learning, and pretraining, covers audio-visual and LLM-based extensions, and discusses long-form processing via segmentation and hypothesis stitching. It also provides a comparative table on AMI, LibriSpeechMix, and LibriMix and draws two main empirical conclusions: no architecture class consistently outperforms the other, and real-world benchmark progress on AMI has stagnated since 2021.

Significance. The survey fills a real gap: there is no recent comprehensive review of end-to-end multi-speaker ASR, and the proposed SIMO/SISO organization is a useful way to structure the literature. The descriptive summaries of individual systems appear mostly accurate, and the paper is appropriately cautious about the limited availability of open-source implementations and the difficulty of fair comparison. However, the central empirical observation about stagnation on AMI rests on Table 2, which mixes evaluation metrics and input granularities. Because the survey's comparative claims are a major part of its contribution, these issues must be addressed before the paper can be relied upon as a reference for the state of the field.

major comments (4)
  1. [Section 6.1 / Table 2] The claim that 'the currently best performance on AMI comes from ... 2021 ... This suggests a stagnation in real-world benchmark progress' is not supported by the table as presented. The Conformer AED entry from Kanda et al. (2021) is evaluated on pre-segmented utterance groups, while later long-form entries such as Transcribe-to-Diarize and SLIDAR must also perform segmentation and speaker-consistent stitching. Comparing these numbers directly conflates recognition quality with the additional difficulty of long-form processing. The comparison should be restricted to a single input granularity, or the stagnation claim should be explicitly qualified as applying only to utterance-group systems.
  2. [Table 2 / Section 3] The taxonomy is applied inconsistently in Table 2. WavLM/wTSE&JSM (Huang et al., 2023) is labeled 'Hybrid' in the AMI panel but 'SIMO' in the LibriMix panel, and the term 'Hybrid' is never defined in Section 3, which presents SIMO and SISO as the two architectural frameworks. Please define 'Hybrid' or remove the category, and ensure that each method receives one consistent classification across all datasets.
  3. [Section 6.1 / Table 2] The table mixes cpWER with WER and SA-WER, and although the footnotes state that marked values cannot be directly compared, the prose in Section 6.1 nevertheless uses them in cross-method comparisons. In particular, the claim that SISO methods outperform SIMO methods on LibriMix compares the b-marked SA-WER of WavLM/wTSE&JSM (10.7b) with unmarked cpWER values from Whisper-SS-TTI and W2V-Sidecar. This is an invalid comparison. The cross-method conclusions should be based only on rows using the same metric, or the prose should acknowledge which rows are being excluded.
  4. [Section 6.1] The statement that 'cpWER has not shown consistent improvement throughout the six-year development period' is stronger than the evidence in Table 2 supports. The table contains a small, non-exhaustive subset of systems, the entries differ in training data, speaker enrollment, and evaluation settings, and the paper itself notes the lack of standardized benchmarks. A lack of improvement among the included methods is not equivalent to a lack of progress in the field. The conclusion should be softened or supported by a more complete and carefully controlled comparison.
minor comments (8)
  1. [Table 2] The third panel heading reads 'LibrMix' and should be 'LibriMix'.
  2. [Section 3.1.2.3] There is a typo in 'mult-speaker ASR'; it should be 'multi-speaker ASR'.
  3. [Section 3.2.2.1] The method name 'Speaker-Disguishable CTC' should be 'Speaker-Distinguishable CTC', matching the description in the text.
  4. [Section 3.2.2.2] The word 'contextural' should be 'contextual' in the sentence about Fan et al. (2024) enhancing the speaker contextual relationship.
  5. [Table 2] The capitalization of 'Utterance Group' is inconsistent across the table rows; some rows use 'Utterance group' and others 'Utterance Group'. Please standardize.
  6. [Table 2] The rows for Hypothesis stitcher (Chang et al., 2021) and Hypothesis clustering (Kashiwagi et al., 2024) are components or processing strategies rather than complete end-to-end ASR systems; the table caption should clarify this so readers do not mistake them for full systems.
  7. [References] Reference formatting is inconsistent, with some entries giving arXiv identifiers and others only the word 'ArXiv' or incomplete venue information; please unify the reference style.
  8. [Section 6.1] The abbreviation 'Params Tr/To' is explained in the table caption, but the text of Section 6.1 also uses it without restating the meaning; consider spelling out 'trainable/total parameters' at first use in the prose.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the survey's synthesis and benchmark comparisons are drawn from external primary sources, with no fitted-input-as-prediction or self-citation chain doing load-bearing work.

full rationale

This paper is a literature survey, not a derivation or empirical prediction. Its central claims are taxonomic observations (SIMO vs. SISO, pre-segmented vs. long-form) and comparative statements about published results, all of which are explicitly attributed to external papers. The performance numbers in Table 2 are reported as taken directly from the original publications ('All results are reported directly from the original papers'), so none of the quantitative observations are produced by fitting a model within this manuscript and then renaming the fit as a prediction. The only self-citations (Li et al. 2023a; Trinh et al. 2024, 2025) are contextual mentions of prior work on diarization and LLM-based text conditioning; they are not invoked to justify the paper's comparative conclusions, and no uniqueness theorem or load-bearing prior result by the authors is used to force the taxonomy. The paper does contain a legitimate comparability caveat: Table 2 mixes cpWER, WER, and SA-WER (footnotes a and b) and mixes utterance-group and long-form protocols, and the stagnation claim about AMI depends on comparing those settings. However, mixing evaluation protocols is a correctness or validity concern about benchmark comparison, not circularity: the manuscript does not define the benchmark results in terms of its own conclusions, nor does it fit parameters to the data it then 'predicts.' Accordingly, under the hard rules requiring a quoted reduction or a fitted-input-called-prediction step, no circular step can be identified, and the appropriate score is 0.

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

The paper introduces no new free parameters, entities, forces, or conserved quantities. All technical content is drawn from cited literature. The ledger entries above are the unstated premises the survey's conclusions rely on: trustworthy reported numbers, metric comparability, and a consistent taxonomy.

assumptions (3)
  • domain assumption Reported benchmark results in the cited papers are accurate and were transcribed faithfully into Table 2.
    All quantitative conclusions in Section 6.1 depend on these numbers; the paper does not re-run or verify them.
  • domain assumption Comparing cpWER, WER, and SA-WER results across methods is valid for drawing qualitative conclusions.
    Section 6.1 adopts cpWER as the primary metric but several cited systems report only WER or SA-WER, and the paper notes these are not directly comparable yet still draws cross-method conclusions.
  • ad hoc to paper The SIMO/SISO/Hybrid trichotomy exhaustively and consistently classifies current E2E multi-speaker ASR systems.
    The taxonomy is the paper's central organizing contribution, yet Table 2 labels the same system (Huang et al. 2023) as Hybrid in the AMI panel and SIMO in the LibriMix panel, making the classification rule ambiguous.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Survey of End-to-End Multi-Speaker Automatic Speech Recognition for Monaural Audio." pith.science (2026). https://pith.science/paper/ZDGYNS6Q

@misc{pith2026250510975,
  author       = {Pith},
  title        = {Pith review of: Survey of End-to-End Multi-Speaker Automatic Speech Recognition for Monaural Audio},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZDGYNS6Q}},
  note         = {Machine review of arXiv:2505.10975}
}
read the original abstract

Monaural multi-speaker automatic speech recognition (ASR) remains challenging due to data scarcity and the intrinsic difficulty of recognizing and attributing words to individual speakers, particularly in overlapping speech. Recent advances have driven the shift from cascade systems to end-to-end (E2E) architectures, which reduce error propagation and better exploit the synergy between speech content and speaker identity. Despite rapid progress in E2E multi-speaker ASR, the field lacks a comprehensive review of recent developments. This survey provides a systematic taxonomy of E2E neural approaches for multi-speaker ASR, highlighting recent advances and comparative analysis. Specifically, we analyze: (1) architectural paradigms (SIMO vs.~SISO) for pre-segmented audio, analyzing their distinct characteristics and trade-offs; (2) recent architectural and algorithmic improvements based on these two paradigms; (3) extensions to long-form speech, including segmentation strategy and speaker-consistent hypothesis stitching. Further, we (4) evaluate and compare methods across standard benchmarks. We conclude with a discussion of open challenges and future research directions towards building robust and scalable multi-speaker ASR.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Diarization-Aware Multi-Speaker Automatic Speech Recognition via Large Language Models

    eess.AS 2025-06 conditional novelty 6.0 of 10

    An LLM conditioned on speaker embeddings and utterance time boundaries jointly transcribes and timestamps overlapping multi-speaker speech.

  2. The DKU System for Multi-Speaker Automatic Speech Recognition in MLC-SLM Challenge

    eess.AS 2025-07 conditional novelty 4.0 of 10

    A challenge system combining speaker diarization, speaker embeddings, and a Qwen2.5 LLM adapter architecture reports 18.08% tcpWER on multilingual multi-speaker ASR, far below the 60.39% baseline.

Reference graph

Works this paper leans on

49 extracted references · 13 canonical work pages · cited by 2 Pith papers

  1. [2]

    Automatic speech recognition: A review. Int. J. Comput. Appl. http://dx.doi.org/10.5120/9722-4190. Baevski, A., Zhou, H., Mohamed, A., Auli, M.,

    work page doi:10.5120/9722-4190

  2. [7]

    In: ICASSP

    End-to-end multi-speaker speech recognition with transformer. In: ICASSP. http://dx.doi.org/10. 1109/ICASSP40776.2020.9054029. Chen, Z., Yoshioka, T., Lu, L., Zhou, T., Meng, Z., Luo, Y., Wu, J., Li, J.,

    arXiv 2020

  3. [8]

    arXiv:2001.11482

    Continuous speech separation: dataset and analysis. arXiv:2001.11482. Chiu, C.-C., Han, W., Zhang, Y., Pang, R., Kishchenko, S., Nguyen, P., Narayanan, A., Liao, H., Zhang, S., Kannan, A., Prabhavalkar, R., Chen, Z., Sainath, T.N., Wu, Y.,

    arXiv 2001

  4. [9]

    arXiv:2005.11262

    LibriMix: An open-source dataset for generalizable speech separation. arXiv:2005.11262. Denisov, P., Vu, N.T.,

    arXiv 2005

  5. [10]

    arXiv:2005.08100

    Conformer: Convolution-augmented transformer for speech recognition. arXiv:2005.08100. He, J., Sawada, N., Miyazaki, K., Toda, T.,

    arXiv 2005

  6. [14]

    arXiv:2106.07447

    HuBERT: Self-supervised speech representation learning by masked prediction of hidden units. arXiv:2106.07447. Hu, E.J., Shen, y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W.,

    arXiv

  7. [18]

    Computer Speech & Language 99 (2026) 101925 16 X

    Large-scale pre-training of end-to-end multi-talker ASR for meeting transcription with single distant microphone. Computer Speech & Language 99 (2026) 101925 16 X. He and J. Whitehill Kang, J., Meng, L., Cui, M., Guo, H., Wu, X., Liu, X., Meng, H.,

    work page 2026

  8. [20]

    Hypothesis Clustering and Merging: Novel MultiTalker Speech Recognition with Speaker Tokens

    Hypothesis clustering and merging: Novel MultiTalker speech recognition with speaker tokens. http://dx.doi.org/10.48550/arXiv.2409.15732. Kolbæk, M., Yu, D., Tan, Z.-H., Jensen, J.H.,

    work page Pith review arXiv doi:10.48550/arxiv.2409.15732

Show all 49 references
  1. [21]

    http://dx.doi.org/10.48550/arXiv.2012.14952

    Bayesian HMM clustering of x-vector sequences (VBx) in speaker diarization: theory, implementation and analysis on standard tasks. http://dx.doi.org/10.48550/arXiv.2012.14952. Latif, S., Shoukat, M., Shamshad, F., Usama, M., Cuayáhuitl, H., Schuller, B.,

    2012 doi arXiv

  2. [22]

    http://dx.doi.org/10

    Sparks of large audio models: A survey and outlook. http://dx.doi.org/10. 48550/arXiv.2308.12792. Li, Z., He, X., Whitehill, J., 2023a. Compositional clustering: Applications to multi-label object recognition and speaker identification. Pattern Recognit. http://dx.doi.org/10.1...

    doi arXiv

  3. [23]

    In: ICASSP

    Compositional embedding models for speaker identification and diarization with simultaneous speech from 2+ speakers. In: ICASSP. http://dx.doi.org/10.1109/ICASSP39728.2021.9413752. Liang, Y., Yu, F., Li, Y., Guo, P., Zhang, S., Chen, Q., Xie, L.,

    2021

  4. [24]

    In: ISCSLP

    Separate-to-recognize: Joint multi-target speech separation and speech recognition for speaker-attributed ASR. In: ISCSLP. http://dx.doi.org/10.1109/ISCSLP57327.2022.10037902. Lu, L., Kanda, N., Li, J., Gong, Y.,

    2022

  5. [25]

    arXiv:1910.06379

    Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separation. arXiv:1910.06379. Luo, Y., Mesgarani, N.,

    1910 arXiv

  6. [28]

    1838–1842

    pp. 1838–1842. http://dx.doi.org/10.21437/Interspeech.2025-

    2025 doi

  7. [29]

    Multimedia Tools Appl

    Automatic speech recognition: a survey. Multimedia Tools Appl. http://dx.doi.org/10.1007/s11042- 020-10073-7. Mao, H.H., Li, S., McAuley, J., Cottrell, G.,

    doi

  8. [30]

    In: Interspeech

    Unified autoregressive modeling for joint end-to-end multi-talker overlapped speech recognition and speaker attribute estimation. In: Interspeech. Meng, L., Hu, S., Kang, J., Li, Z., Wang, Y., Wu, W., Wu, X., Liu, X., Meng, H., 2024a. Large language model can transcribe speech...

    arXiv

  9. [31]

    In: ICASSP

    A sidecar separator can convert a single-talker speech recognition system to a multi-talker one. In: ICASSP. http://dx.doi.org/10.1109/ICASSP49357.2023.10095295. Meng, L., Kang, J., Wang, Y., Jin, Z., Wu, X., Liu, X., Meng, H., 2024b. Empowering whisper as a joint multi-talker...

    2023

  10. [32]

    http://dx.doi.org/10.48550/arXiv.2409.20301

    Alignment-free training for transducer-based multi-talker ASR. http://dx.doi.org/10.48550/arXiv.2409.20301. Park, T.J., Medennikov, I., Dhawan, K., Wang, W., Huang, H., Koluguri, N.R., Puvvada, K.C., Balam, J., Ginsburg, B.,

    doi arXiv

  11. [34]

    IEEE/ACM Trans

    End-to-end speech recognition: A survey. IEEE/ACM Trans. Audio Speech Lang. Process. http://dx.doi.org/10.1109/TASLP.2023.3328283. Qian, Y.-m., Weng, C., Chang, X.-k., Wang, S., Yu, D.,

    2023

  12. [36]

    arXiv preprint arXiv:2106.04624

    SpeechBrain: A general-purpose speech toolkit. arXiv preprint arXiv:2106.04624. Rose, R., Chang, O., Siohan, O.,

    arXiv

  13. [37]

    arXiv:2306.16398

    Cascaded encoders for fine-tuning ASR models on overlapped speech. arXiv:2306.16398. Sakuma, A., Sato, H., Sugano, R., Kumano, T., Kawai, Y., Ogawa, T.,

    arXiv

  14. [38]

    In: Proceedings of the Annual Meeting of the Assoc

    A purely end-to-end system for multi-speaker speech recognition. In: Proceedings of the Annual Meeting of the Assoc. for Comp. Linguistics. http://dx.doi.org/10.18653/v1/P18-1244. Sell, G., Garcia-Romero, D.,

    doi

  15. [39]

    In: ICASSP

    End-to-end multi-speaker speech recognition. In: ICASSP. http://dx.doi.org/10.1109/ICASSP. 2018.8461893. Computer Speech & Language 99 (2026) 101925 17 X. He and J. Whitehill Shakeel, M., Sudo, Y., Peng, Y., Lin, C.-J., Watanabe, S.,

    2026

  16. [40]

    In: Interspeech

    CASA-ASR: Context-aware speaker-attributed ASR. In: Interspeech. http: //dx.doi.org/10.21437/Interspeech.2023-601. Shi, H., Fujita, Y., Mizumoto, T., Liu, L., Kojima, A., Sudo, Y.,

    2023 doi

  17. [41]

    arXiv:2509.04488

    Serialized output prompting for large language model-based multi-talker speech recognition. arXiv:2509.04488. URL https://arxiv.org/abs/2509.04488. Shi, H., Gao, Y., Ni, Z., Kawahara, T., 2024a. Serialized speech information guidance with overlapped encoding separation for mul...

    arXiv

  18. [42]

    arXiv:2506.10779

    Improving named entity transcription with contextual LLM-based revision. arXiv:2506.10779. URL https://arxiv.org/abs/ 2506.10779. Trinh, V.A., Southwell, R., Guan, Y., He, X., Wang, Z., Whitehill, J.,

    arXiv

  19. [43]

    arXiv:2406.06582

    Discrete multimodal transformers with a pretrained large language model for mixed-supervision speech processing. arXiv:2406.06582. URL https://arxiv.org/abs/2406.06582. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., Polosukhin, I.,

    arXiv

  20. [44]

    arXiv:1706.03762

    Attention is all you need. arXiv:1706.03762. von Neumann, T., Boeddeker, C., Delcroix, M., Haeb-Umbach, R.,

    arXiv

  21. [45]

    http://dx.doi.org/10.48550/arXiv.2307.11394

    MeetEval: A toolkit for computation of word error rates for meeting transcription systems. http://dx.doi.org/10.48550/arXiv.2307.11394. von Neumann, T., Boeddeker, C., Drude, L., Kinoshita, K., Delcroix, M., Nakatani, T., Haeb-Umbach, R.,

    doi arXiv

  22. [46]

    http://dx.doi.org/10.1109/SLT61566.2024.10832215

    Resource-efficient adaptation of speech foundation models for multi-speaker ASR. http://dx.doi.org/10.1109/SLT61566.2024.10832215. Wang, Z.-Q., Roux, J.L., Hershey, J.R.,

    2024

  23. [47]

    In: ICASSP

    Alternative objective functions for deep clustering. In: ICASSP. http://dx.doi.org/10.1109/ICASSP.2018.8462507. Wang, J., Wang, W., Dhawan, K., Park, T., Kim, M., Medennikov, I., Huang, H., Koluguri, N., Balam, J., Ginsburg, B.,

    2018

  24. [48]

    Watanabe, S., Mandel, M.I., Barker, J., Vincent, E.,

    http://dx.doi.org/10.1109/ICASSP49660.2025.10889841. Watanabe, S., Mandel, M.I., Barker, J., Vincent, E.,

    2025

  25. [49]

    arXiv: 2004.09249

    CHiME-6 challenge: Tackling multispeaker speech recognition for unsegmented recordings. arXiv: 2004.09249. Watanabe, S., et al.,

    2004 arXiv

  26. [50]

    3021–3025

    pp. 3021–3025. http://dx.doi.org/10.21437/Interspeech.2021-2128. Yang, M., Kanda, N., Wang, X., Wu, J., Sivasankaran, S., Chen, Z., Li, J., Yoshioka, T.,

    2021 doi

  27. [51]

    In: ICASSP

    Simulating realistic speech overlaps improves multi-talker ASR. In: ICASSP. Yu, D., Chang, X., Qian, Y., 2017a. Recognizing multi-talker speech with permutation invariant training. In: Interspeech. Yu, F., Du, Z., Zhang, S., Lin, Y., Xie, L., 2022a. A comparative study on spea...

  28. [52]

    Zhang, W., Chang, X., Qian, Y., Watanabe, S.,

    http://dx.doi.org/10.21437/ Interspeech.2019-3192. Zhang, W., Chang, X., Qian, Y., Watanabe, S.,

    2019

  29. [53]

    IEEE/ACM Trans

    Improving end-to-end single-channel multi-talker speech recognition. IEEE/ACM Trans. Audio Speech Lang. Process. http://dx.doi.org/10.1109/TASLP.2020.2988423. Zheng, L., Zhu, H., Tian, S., Zhao, Q., Li, T.,

    2020

  30. [54]

    IEEE Signal Process

    Unsupervised domain adaptation on end-to-end multi-talker overlapped speech recognition. IEEE Signal Process. Lett. http://dx.doi.org/10.1109/LSP.2024.3487795. Computer Speech & Language 99 (2026) 101925 18

    2026

  31. [2010]

    Superhuman multi-talker speech recognition: A graphical modeling approach. Comput. Speech Lang. http://dx.doi.org/10.1016/j.csl.2008.11.001. Hsu, W., Bolte, B., Tsai, Y.H., Lakhotia, K., Salakhutdinov, R., Mohamed, A.,

    2008 doi

  32. [2012]

    IEEE Trans

    Speaker diarization: A review of recent research. IEEE Trans. Audio Speech Lang. Process. http://dx.doi.org/10.1109/TASL.2011.2125954. Arora, S., Singh, R.,

    2011

  33. [2016]

    In: ICASSP

    Deep clustering: Discriminative embeddings for segmentation and separation. In: ICASSP. http: //dx.doi.org/10.1109/ICASSP.2016.7471631. Hershey, J., Rennie, S., Olsen, P., Kristjansson, T.,

    2016

  34. [2017]

    arXiv:1711.00541

    TasNet: time-domain audio separation network for real-time, single-channel speech separation. arXiv:1711.00541. Ma, P., Haliassos, A., Fernandez-Lopez, A., Chen, H., Petridis, S., Pantic, M.,

    arXiv

  35. [2018]

    Past review, current progress, and challenges ahead on the cocktail party problem. Front. Info. Technol. Electron. Eng. http://dx.doi.org/10.1631/FITEE.1700814. Radford, A., Kim, J.W., Xu, T., Brockman, G., McLeavey, C., Sutskever, I.,

    doi

  36. [2019]

    In: ICASSP

    End-to-end monaural multi-speaker ASR system without pretraining. In: ICASSP. http://dx.doi.org/10.1109/ ICASSP.2019.8682822. Chang, X., Zhang, W., Qian, Y., Roux, J.L., Watanabe, S.,

    2019

  37. [2020]

    arXiv:2006.11477

    wav2vec 2.0: A framework for self-supervised learning of speech representations. arXiv:2006.11477. Braga, O., Makino, T., Siohan, O., Liao, H.,

    2006 arXiv

  38. [2022]

    In: ICASSP

    Transcribe-to-diarize: Neural speaker diarization for unlimited number of speakers using end-to-end speaker-attributed ASR. In: ICASSP. Kanda, N., Xiao, X., Wu, J., Zhou, T., Gaur, Y., Wang, X., Meng, Z., Chen, Z., Yoshioka, T., 2021c. A comparative study of modular and joint ...

  39. [2024]

    http://dx.doi.org/10.21437/Interspeech.2024-90

    LibriheavyMix: A 20,000-hour dataset for single-channel reverberant multi-talker speech separation, ASR and speaker diarization. http://dx.doi.org/10.21437/Interspeech.2024-90. Kanda, N., Chang, X., Gaur, Y., Wang, X., Meng, Z., Chen, Z., Yoshioka, T., 2021a. Investigation of ...

    2024

  40. [2025]

    CMT-LLM: Contextual multi-talker ASR utilizing large language models. pp. 2575–2579. http://dx.doi.org/10. 21437/Interspeech.2025-943. Hershey, J.R., Chen, Z., Le Roux, J., Watanabe, S.,

    2025

  41. [2026]

    DiCoW: Diarization- conditioned whisper for target speaker automatic speech recognition. Comput. Speech Lang. 95, 101841. http://dx.doi.org/10.1016/j.csl.2025.101841, URL https://www.sciencedirect.com/science/article/pii/S088523082500066X. Prabhavalkar, R., Hori, T., Sainath, ...

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.