Pith. sign in

REVIEW 5 major objections 4 minor 41 references

TS-SUPERB: A Target Speech Processing Benchmark for Speech Self-Supervised Learning Models

T0 review · 5 major / 4 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 TS-SUPERB, a new benchmark of four target-speaker tasks, shows that SSL model rankings change when a model must extract one voice from a mixture, and that jointly training these tasks through one shared encoder improves performance.

desk verdict Useful target-speaker SSL benchmark, but the main-table numbers don't match the released-code numbers, so the rankings are not yet canonical. read the letter →

arxiv 2505.06660 v1 pith:57KQXBIS submitted 2025-05-10 cs.CL cs.SDeess.AS

Junyi Peng , Takanori Ashihara , Marc Delcroix , Tsubasa Ochiai , Oldrich Plchot , Shoko Araki , Jan Černocký This is my paper · ORCID
classification cs.CLcs.SDeess.AS
keywords self-supervisedlearningtargetspeechextractionpersonalizedenhancementtarget-speakerASRvoiceactivitydetectionbenchmarkmulti-talkermulti-task
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 introduces TS-SUPERB, a benchmark for target-speaker speech processing that evaluates self-supervised speech models on four tasks: target speech extraction, personalized speech enhancement, target-speaker ASR, and personalized voice activity detection, all on two-speaker mixtures. The central claim is that a model's performance on ordinary single-speaker tasks does not reveal how well it can lock onto one speaker's voice in a noisy mixture, so separate evaluation is needed. Across seven SSL models, no single model wins all four tasks, and WavLM variants tend to lead on denoising; TS-ASR correlates only weakly with plain ASR. The paper also shows that jointly training pairs of target-speaker tasks through a shared target-speech encoder improves several metrics, suggesting the tasks share usable structure.

What carries the argument

The central object is the shared target speech encoder: a speaker encoder that uses multi-head factorized attention pooling to turn an enrollment utterance into a speaker vector, together with an extractor made of two bidirectional LSTM layers that fuses that vector into SSL features of the mixture through broadcast multiplication. This encoder is what lets all four tasks condition on 'who to listen to,' and because its parameters are shared across tasks, it is also what makes multi-task joint training possible. By keeping the encoder architecture fixed and the SSL model frozen by default, the benchmark isolates the contribution of each SSL model to target-speaker performance.

What would settle it

Run the four TS tasks with a stronger decoder (e.g., deeper extractor or full fine-tuning of each SSL encoder); if any model's relative ranking changes on TSE, PSE, TS-ASR, or PVAD, the benchmark's comparative claim is an artifact of the fixed small downstream network rather than a property of the SSL models.

Watch

Extended reading notes

Core claim

TS-SUPERB provides a standardized comparison showing that SSL model strength on target-speaker tasks is not inferable from single-speaker benchmarks, because every TS task combines speaker identification with content or acoustic extraction. The paper reports that all four TS tasks correlate strongly with speech separation, while TS-ASR and plain ASR correlate only weakly, and that WavLM models trained with mixed-speaker augmentation lead on denoising tasks. Using a unified architecture—an SSL-based speaker encoder with multi-head factorized attention pooling over enrollment speech, plus an SSL-based extractor of two BLSTM layers that fuses the speaker vector into mixture features—the paper finds that jointly optimizing TSE with TS-ASR and PSE with PVAD improves several metrics over single-task training, with additional gains when the SSL encoder is fine-tuned. An appendix notes that the released codebase yields slightly different numbers than the main tables, so the benchmark's contribution is the evaluation structure and qualitative findings rather than exact scores.

Load-bearing premise

The rankings hold only if the simple, frozen downstream network used for every model is powerful enough to expose each SSL model's real strengths on target-speaker tasks rather than hiding them behind the decoder.

Editorial extensions

If this is right

  • SSL model rankings on target-speaker tasks cannot be read off from single-speaker SUPERB-style results, so TS-SUPERB adds a separate evaluation axis for speech SSL models.
  • Models pretrained with denoising or multi-speaker augmentation, such as the WavLM variants, tend to lead on TSE and PSE, indicating that pretraining data composition shapes target-speaker ability.
  • TS-ASR requires both lower-layer speaker-related information and upper-layer semantic information, so layer-weighted SSL features for TS-ASR should combine bottom and top layers rather than relying on one region.
  • Joint training can improve denoising and detection metrics while slightly worsening TS-ASR word error rate, and fine-tuning the SSL encoder recovers further gains.
  • Because all TS tasks correlate strongly with speech separation, progress on separation is a meaningful proxy for progress across target-speaker tasks.

Reading between the lines

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

  • If the shared-encoder design scales beyond the two pairwise setups tested, a single SSL backbone could simultaneously enhance, transcribe, and detect one target voice in a device; training all four tasks together is a natural next experiment.
  • The layer-weight pattern suggests a cheap recipe: per-task layer weighting or sub-selecting a bottom-heavy set of SSL layers could improve TS task performance without fine-tuning the SSL model.
  • PVAD's weak correlation with speaker verification implies it may not need fine-grained speaker embeddings, so a lighter or frame-level speaker condition deserves a direct test.
  • Extending the benchmark to unseen overlap ratios, more than two speakers, or real meeting recordings would show whether the model rankings persist outside LibriMix conditions.
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 / 4 minor

Summary. The manuscript proposes TS-SUPERB, a benchmark that evaluates seven speech self-supervised learning (SSL) models on four target-speaker tasks: target speech extraction (TSE), personalized speech enhancement (PSE), personalized voice activity detection (PVAD), and target-speaker ASR (TS-ASR). It introduces a unified downstream architecture composed of an SSL-based target speech encoder and task-specific decoders, with SSL models kept frozen by default, and reports results across the four tasks, layer-wise weight analyses, Spearman correlations with related single-speaker tasks, and multi-task learning experiments (TSE+TS-ASR and PSE+PVAD). The paper's central claims are that target-speaker performance cannot be inferred from single-speaker SUPERB tasks and that joint training of target-speech encoders across TS tasks can improve performance. Appendix A reports results from a refactored released codebase that differ from the main-table results.

Significance. If the empirical results were stable, TS-SUPERB would be a valuable contribution: it fills a clear gap in SSL benchmarking by targeting multi-talker scenarios, uses publicly available datasets, releases code, and proposes a shared architecture that makes multi-task analysis possible. The layer-wise and correlation analyses are interesting, and the multi-task experiments provide useful initial evidence for joint training. The benchmark is not circular in its construction: the four tasks use external datasets and objective metrics, and the SSL features are evaluated as frozen upstream representations. However, the empirical confidence is substantially weakened by the discrepancy between Table II and Appendix A, by the absence of repeated-run statistics, and by the heterogeneous provenance of the reference task scores; these issues must be resolved before the comparative claims can be relied upon.

major comments (5)
  1. [Table II vs Appendix A] The main experimental table is not reproducible from the released-code results reported in the paper itself. For PSE average SI-SDRi, HuBERT Base changes from 10.36 (Table II) to 8.61 (Table VI), WavLM Base changes from 11.01 to 9.65, and WavLM Base+ changes from 10.96 to 10.01, which reverses the Base/Base+ ordering shown in Table II. TS-ASR WER without LM also changes substantially, e.g., HuBERT Base from 41.75 to 36.86 and WavLM Base+ from 29.09 to 24.75. Because Appendix A reports only three of the seven models, the full model rankings, layer-weight analyses (Fig. 2), and Spearman correlations (Fig. 3) are not reproducible from the released code. The authors must either replace Table II and all derived analyses with the refactored-code numbers or explain why the Table II numbers remain canonical; as written, the central claims rest on numbers the paper itself identifies as outdated.
  2. [Section IV-A, Tables II-IV] All reported results appear to be single runs, with no error bars, confidence intervals, or significance tests. Several conclusions in Section IV-A hinge on small margins, such as the PVAD mAP of 0.951 for WavLM Base versus 0.945 for data2vec Base, and the Spearman correlations in Fig. 3 are computed over only seven models. The authors should report multiple seeds (or otherwise quantify uncertainty) and restrict their ranking claims to differences that are actually distinguishable.
  3. [Table II, Fig. 3] The Sep, ASR, and SV columns in Table II are taken from references [10] and [37] and therefore are not produced under the same decoder, training recipe, and data protocol as the TS-SUPERB columns; yet they are pooled into the Spearman correlations in Fig. 3. The resulting statements, such as "all TS-SUPERB tasks exhibit a strong correlation with Sep" or "TS-ASR has a surprisingly weak correlation with ASR," are vulnerable to protocol artifacts. The correlation analysis should either be restricted to internally consistent measurements or the reference tasks should be rerun under the same protocol.
  4. [Section IV-B, Fig. 3] The main claim that target-speaker performance "cannot be easily inferred" from single-speaker tasks is not directly supported by the evidence in Fig. 3, which reports strong positive correlations with Sep and SV for most TS-SUPERB tasks. To make the claim operational, the authors should specify what inference procedure they have in mind (e.g., predicting one metric from another, or ranking models) and demonstrate quantitatively where it fails; a Spearman correlation coefficient alone does not establish non-inferability, and some of the reported high correlations point in the opposite direction.
  5. [Section III-B] The benchmark's ranking results depend on the fixed frozen-SSL target speech encoder used across tasks. The manuscript does not test whether this particular BLSTM-based encoder and the frozen-feature constraint distort differences among SSL models; the only fine-tuning experiment is for one upstream model in Tables III and IV. If the decoder capacity or the freezing policy masks or amplifies model differences, the comparative conclusions may not generalize. A useful control would be to re-rank a subset of upstream models under an alternative decoder or with full fine-tuning.
minor comments (4)
  1. [Throughout] The term "PV AD" is written with a space throughout; it should be "PVAD" for consistency with the task name used elsewhere.
  2. [Section V] In the conclusion, "Libr2Mix" should be "Libri2Mix."
  3. [Appendix A] The appendix states that the refactored-code results "slightly differ" from Table II, but differences such as 1.75 dB in PSE SI-SDRi and 7 points in TS-ASR WER are substantial; the wording should be adjusted to reflect the magnitude.
  4. [Fig. 2] The layer-weight distribution is shown only for WavLM Base+; the caption should state this explicitly and explain the notation for the "0th layer."

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TS-SUPERB's benchmark results are empirical, externally measured, and not forced by construction.

full rationale

The paper's central claims are that SSL model performance on target-speaker tasks cannot be inferred from single-speaker SUPERB tasks, and that joint training across TS tasks can help. Both claims are supported by measurements on external datasets (Libri2Mix and noisy SparseLibri2Mix) with standard metrics (SI-SDRi, WER, mAP, etc.). The Spearman correlation analysis in Fig. 3 is a post-hoc statistical summary of those measured scores, not a quantity defined by the benchmark's architecture. The multi-task results in Tables III and IV compare the same architecture with and without a multi-task loss, so the improvement is an empirical outcome rather than an algebraic identity. Self-citations appear as components (MHFA speaker encoder from [7], target extraction module mechanisms from [16]) and as external comparison scores for Sep/SV from [37]. These cited numbers are externally measured values, not fitted parameters of this paper, and the conclusions do not reduce to the citations. The Appendix A disclosure that the released, refactored code produces different numbers (e.g., PSE SI-SDRi HuBERT Base 10.36 to 8.61, TS-ASR w/o LM WavLM Base+ 29.09 to 24.75) is a reproducibility limitation, not circularity: it does not show that any prediction is equivalent to an input by definition. No specific equation or construction was found that makes the claimed results equal to their inputs, so under the hard rules the appropriate score is 0.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The benchmark's central evaluation is empirical, so the ledger mostly records assumptions about dataset representativeness, decoder sufficiency, and statistical power. The only hand-chosen values that affect the multi-task and PVAD results are alpha equals 0.5 and the PVAD decoder width.

free parameters (2)
  • Multi-task loss weight alpha = 0.5
    Set to 0.5 for all joint-training experiments (Section III-G); no sweep or tuning is reported, and the MTL comparison relies on this single value.
  • PVAD BLSTM hidden dimension = 32
    The PVAD decoder uses 32 units while other tasks use 512, chosen based on preliminary PVAD experiments that showed negligible differences (Section III-B).
assumptions (3)
  • domain assumption Libri2Mix and noisy SparseLibri2Mix are representative of target-speaker processing in multi-talker conditions.
    The benchmark's external validity as a general evaluation rests on these public datasets capturing the difficulty of TS tasks (Section III-A).
  • domain assumption A lightweight BLSTM decoder operating on frozen SSL features is sufficient to rank SSL models for TS tasks.
    The design follows SUPERB's simple-decoder philosophy; if the decoder is too weak or too strong, the model rankings may not reflect SSL quality (Section III-B).
  • domain assumption Spearman correlations computed from seven SSL models are meaningful indicators of task relationships.
    Figure 3 is based on only seven data points with no significance testing or confidence intervals (Section IV-B).

how reviews work

0 comments
Cite this review

Pith. "Pith review of TS-SUPERB: A Target Speech Processing Benchmark for Speech Self-Supervised Learning Models." pith.science (2026). https://pith.science/paper/57KQXBIS

@misc{pith2026250506660,
  author       = {Pith},
  title        = {Pith review of: TS-SUPERB: A Target Speech Processing Benchmark for Speech Self-Supervised Learning Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/57KQXBIS}},
  note         = {Machine review of arXiv:2505.06660}
}
read the original abstract

Self-supervised learning (SSL) models have significantly advanced speech processing tasks, and several benchmarks have been proposed to validate their effectiveness. However, previous benchmarks have primarily focused on single-speaker scenarios, with less exploration of target-speaker tasks in noisy, multi-talker conditions -- a more challenging yet practical case. In this paper, we introduce the Target-Speaker Speech Processing Universal Performance Benchmark (TS-SUPERB), which includes four widely recognized target-speaker processing tasks that require identifying the target speaker and extracting information from the speech mixture. In our benchmark, the speaker embedding extracted from enrollment speech is used as a clue to condition downstream models. The benchmark result reveals the importance of evaluating SSL models in target speaker scenarios, demonstrating that performance cannot be easily inferred from related single-speaker tasks. Moreover, by using a unified SSL-based target speech encoder, consisting of a speaker encoder and an extractor module, we also investigate joint optimization across TS tasks to leverage mutual information and demonstrate its effectiveness.

Figures

Figures reproduced from arXiv: 2505.06660 by the authors.

Figure 1
Figure 1. Architecture of proposed TS-SUPERB system. All downstream models [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The weight distribution of SSL layers. Note that the 0th layer denotes [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The Spearman’s rank correlation between tasks. [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

41 extracted references · 34 canonical work pages

  1. [10]

    SUPERB: Speech Processing Universal PERformance Benchmark,

    S.-W. Yang, P.-H. Chi, Y .-S. Chuang, C.-I. J. Lai, K. Lakhotia, Y . Y . Lin, A. T. Liu, J. Shi, X. Chang, G.-T. Lin, T.-H. Huang, W.-C. Tseng, K. Lee, D.-R. Liu, Z. Huang, S. Dong, S.-W. Li, S. Watanabe, A. Mohamed, and H. yi Lee, “SUPERB: Speech Processing Universal PERformance Benchmark,” inINTERSPEECH, 2021, pp. 1194–1198

    work page 2021

  2. [37]

    Probing self-supervised learning models with target speech extraction,

    J. Peng, M. Delcroix, T. Ochiai, O. Plchot, T. Ashihara, S. Araki, and J. Cernocky, “Probing self-supervised learning models with target speech extraction,” inICASSP SASB Workshops, 2024

    work page 2024

  3. [1]

    wav2vec 2.0: A framework for self-supervised learning of speech representations,

    A. Baevski, Y . Zhou, A. Mohamed, and M. Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representations,” Advances in neural information processing systems, vol. 33, pp. 12 449– 12 460, 2020

    2020

  4. [2]

    data2vec: A general framework for self-supervised learning in speech, vision and language,

    A. Baevski, W.-N. Hsu, Q. Xu, A. Babu, J. Gu, and M. Auli, “data2vec: A general framework for self-supervised learning in speech, vision and language,” inICML, 2022, pp. 1298–1312

    work page 2022

  5. [3]

    HuBERT: Self-supervised speech representation learning by masked prediction of hidden units,

    W.-N. Hsu, B. Bolte, Y .-H. H. Tsai, K. Lakhotia, R. Salakhutdinov, and A. Mohamed, “HuBERT: Self-supervised speech representation learning by masked prediction of hidden units,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 29, pp. 3451–3460, 2021

    2021

  6. [4]

    WavLM: Large-Scale Self-Supervised Pre- Training for Full Stack Speech Processing,

    S. Chen, C. Wang, Z. Chen, Y . Wu, S. Liu, Z. Chen, J. Li, N. Kanda, T. Yoshioka, X. Xiaoet al., “WavLM: Large-Scale Self-Supervised Pre- Training for Full Stack Speech Processing,”IEEE Journal of Selected Topics in Signal Processing, vol. 16, no. 6, pp. 1505–1518, 2022

    work page 2022

  7. [5]

    Parameter-efficient cross-language transfer learning for a language- modular audiovisual speech recognition,

    Z. Li, T. Graave, J. Liu, T. Lohrenz, S. Kunzmann, and T. Fingscheidt, “Parameter-efficient cross-language transfer learning for a language- modular audiovisual speech recognition,” inASRU, 2023, pp. 1–8

    work page 2023

  8. [6]

    Large-scale self-supervised speech representation learning for automatic speaker verification,

    Z. Chen, S. Chen, Y . Wu, Y . Qian, C. Wang, S. Liu, Y . Qian, and M. Zeng, “Large-scale self-supervised speech representation learning for automatic speaker verification,” inICASSP, 2022, pp. 6147–6151

    work page 2022

Show all 41 references
  1. [7]

    An attention-based backend allowing efficient fine-tuning of transformer models for speaker verification,

    J. Peng, O. Plchot, T. Stafylakis, L. Mo ˇsner, L. Burget, and J. ˇCernock`y, “An attention-based backend allowing efficient fine-tuning of transformer models for speaker verification,” inSLT, 2023, pp. 555–562

    2023

  2. [8]

    Exploring WavLM on speech enhancement,

    H. Song, S. Chen, Z. Chen, Y . Wu, T. Yoshioka, M. Tang, J. W. Shin, and S. Liu, “Exploring WavLM on speech enhancement,” inSLT, 2023, pp. 451–457

    2023

  3. [9]

    Boosting self-supervised embeddings for speech enhancement,

    K.-H. Hung, S. wei Fu, H.-H. Tseng, H.-T. Chiang, Y . Tsao, and C.-W. Lin, “Boosting self-supervised embeddings for speech enhancement,” in INTERSPEECH, 2022, pp. 186–190

    2022

  4. [11]

    SUPERB- SG: Enhanced speech processing universal PERformance benchmark for semantic and generative capabilities,

    H.-S. Tsai, H.-J. Chang, W.-C. Huang, Z. Huang, K. Lakhotia, S.-w. Yang, S. Dong, A. Liu, C.-I. Lai, J. Shi, X. Chang, P. Hall, H.-J. Chen, S.-W. Li, S. Watanabe, A. Mohamed, and H.-y. Lee, “SUPERB- SG: Enhanced speech processing universal PERformance benchmark for semantic an...

    2022

  5. [12]

    ML- SUPERB: Multilingual Speech Universal PERformance Benchmark,

    J. Shi, D. Berrebbi, W. Chen, E.-P. Hu, W.-P. Huang, H.-L. Chung, X. Chang, S.-W. Li, A. Mohamed, H. yi Lee, and S. Watanabe, “ML- SUPERB: Multilingual Speech Universal PERformance Benchmark,” in INTERSPEECH, 2023, pp. 884–888

    2023

  6. [13]

    Min- isuperb: Lightweight benchmark for self-supervised speech models,

    Y .-H. Wang, H.-Y . Chen, K.-W. Chang, W. Hsu, and H.-Y . Lee, “Min- isuperb: Lightweight benchmark for self-supervised speech models,” in ASRU, 2023, pp. 1–8

    2023

  7. [14]

    A V- SUPERB: A multi-task evaluation benchmark for audio-visual represen- tation models,

    Y . Tseng, L. Berry, Y .-T. Chen, I.-H. Chiu, H.-H. Lin, M. Liu, P. Peng, Y .-J. Shih, H.-Y . Wang, H. Wu, P.-Y . Huang, C.-M. Lai, S.-W. Li, D. Harwath, Y . Tsao, A. Mohamed, C.-L. Feng, and H.-Y . Lee, “A V- SUPERB: A multi-task evaluation benchmark for audio-visual represen...

    2024

  8. [15]

    EMO-SUPERB: An in-depth look at speech emotion recognition,

    H. Wu, H.-C. Chou, K.-W. Chang, L. Goncalves, J. Du, J.-S. R. Jang, C.-C. Lee, and H.-Y . Lee, “EMO-SUPERB: An in-depth look at speech emotion recognition,”arXiv preprint arXiv:2402.13018, 2024

    2024 arXiv

  9. [16]

    Neural target speech extraction: An overview,

    K. Zmolikova, M. Delcroix, T. Ochiai, K. Kinoshita, J. ˇCernock´y, and D. Yu, “Neural target speech extraction: An overview,”IEEE Signal Processing Magazine, vol. 40, no. 3, pp. 8–29, 2023

    2023

  10. [17]

    Personalized speech enhancement: New models and comprehensive evaluation,

    S. E. Eskimez, T. Yoshioka, H. Wang, X. Wang, Z. Chen, and X. Huang, “Personalized speech enhancement: New models and comprehensive evaluation,” inICASSP, 2022, pp. 356–360

    2022

  11. [18]

    Investigation of end-to-end speaker-attributed ASR for continuous multi-talker recordings,

    N. Kanda, X. Chang, Y . Gaur, X. Wang, Z. Meng, Z. Chen, and T. Yoshioka, “Investigation of end-to-end speaker-attributed ASR for continuous multi-talker recordings,” inSLT, 2021, pp. 809–816

    2021

  12. [19]

    Personal V AD: Speaker-conditioned voice activity detection,

    S. Ding, Q. Wang, S.-Y . Chang, L. Wan, and I. Lopez Moreno, “Personal V AD: Speaker-conditioned voice activity detection,” inOdyssey, 2020, pp. 433–439

    2020

  13. [20]

    Look once to hear: Target speech hearing with noisy examples,

    B. Veluri, M. Itani, T. Chen, T. Yoshioka, and S. Gollakota, “Look once to hear: Target speech hearing with noisy examples,” inCHI, 2024, pp. 1–16

    2024

  14. [21]

    Streaming target-speaker ASR with neural transducer,

    T. Moriya, H. Sato, T. Ochiai, M. Delcroix, and T. Shinozaki, “Streaming target-speaker ASR with neural transducer,”arXiv preprint arXiv:2209.04175, 2022

    2022 arXiv

  15. [22]

    Target speech extraction with pre-trained self-supervised learning mod- els,

    J. Peng, M. Delcroix, T. Ochiai, O. Plchot, S. Araki, and J. ˇCernock´y, “Target speech extraction with pre-trained self-supervised learning mod- els,” inICASSP, 2024

    2024

  16. [24]

    Improving speaker verification with self-pretrained trans- former models,

    J. Peng, O. Plchot, T. Stafylakis, L. Mosner, L. Burget, and J. H. ˇCernock´y, “Improving speaker verification with self-pretrained trans- former models,” inINTERSPEECH, 2023, pp. 5361–5365

    2023

  17. [25]

    Weakly-supervised speech pre-training: A case study on target speech recognition,

    W. Zhang and Y . Qian, “Weakly-supervised speech pre-training: A case study on target speech recognition,” inINTERSPEECH, 2023, pp. 3517– 3521

    2023

  18. [26]

    Towards learning a universal non-semantic representation of speech,

    J. Shor, A. Jansen, R. Maor, O. Lang, O. Tuval, F. de Chaumont Quitry, M. Tagliasacchi, I. Shavitt, D. Emanuel, and Y . Haviv, “Towards learning a universal non-semantic representation of speech,” inINTERSPEECH, 2020, pp. 140–144

    2020

  19. [27]

    LeBenchmark: A reproducible framework for assessing self-supervised representation learning from speech,

    S. Evain, H. Nguyen, H. Le, M. Z. Boito, S. Mdhaffar, S. Alisamir, Z. Tong, N. Tomashenko, M. Dinarelli, T. Parcollet, A. Allauzen, Y . Est`eve, B. Lecouteux, F. Portet, S. Rossato, F. Ringeval, D. Schwab, and L. Besacier, “LeBenchmark: A reproducible framework for assessing s...

    2021

  20. [28]

    LeBench- mark 2.0: A standardized, replicable and enhanced framework for self-supervised representations of French speech,

    T. Parcollet, H. Nguyen, S. Evain, M. Z. Boito, A. Pupier, S. Mdhaffar, H. Le, S. Alisamir, N. Tomashenko, M. Dinarelliet al., “LeBench- mark 2.0: A standardized, replicable and enhanced framework for self-supervised representations of French speech,”Computer Speech & Language...

    2024

  21. [29]

    HAER: Holis- tic evaluation of audio representations,

    J. Turian, J. Shier, H. R. Khan, B. Raj, B. W. Schuller, C. J. Steinmetz, C. Malloy, G. Tzanetakis, G. Velarde, K. McNallyet al., “HAER: Holis- tic evaluation of audio representations,” inNeurIPS 2021 Competitions and Demonstrations Track, 2022, pp. 125–145

    2021

  22. [30]

    XTREME-S: Evaluating cross-lingual speech representations,

    A. Conneau, A. Bapna, Y . Zhang, M. Ma, P. von Platen, A. Lozhkov, C. Cherry, Y . Jia, C. Rivera, M. Kale, D. van Esch, V . Axelrod, S. Khanuja, J. Clark, O. Firat, M. Auli, S. Ruder, J. Riesa, and M. John- son, “XTREME-S: Evaluating cross-lingual speech representations,” in I...

    2022

  23. [31]

    Unveiling the linguistic capabilities of a self-supervised speech model through cross- lingual benchmark and layer-wise similarity analysis,

    T. Ashihara, M. Delcroix, Y . Ijima, and M. Kashino, “Unveiling the linguistic capabilities of a self-supervised speech model through cross- lingual benchmark and layer-wise similarity analysis,”IEEE Access, vol. 12, pp. 98 835–98 855, 2024

    2024

  24. [32]

    Adapting self- supervised models to multi-talker speech recognition using speaker embeddings,

    Z. Huang, D. Raj, P. Garc ´ıa, and S. Khudanpur, “Adapting self- supervised models to multi-talker speech recognition using speaker embeddings,” inICASSP, 2023, pp. 1–5

    2023

  25. [33]

    LibriMix: An open-source dataset for generalizable speech separation,

    J. Cosentino, M. Pariente, S. Cornell, A. Deleforge, and E. Vincent, “LibriMix: An open-source dataset for generalizable speech separation,” arXiv preprint arXiv:2005.11262, 2020

    2005 arXiv

  26. [34]

    LibriSpeech: An ASR corpus based on public domain audio books,

    V . Panayotov, G. Chen, D. Povey, and S. Khudanpur, “LibriSpeech: An ASR corpus based on public domain audio books,” inICASSP, 2015

    2015

  27. [35]

    WHAM!: Extending speech separation to noisy environments,

    G. Wichern, J. Antognini, M. Flynn, L. R. Zhu, E. McQuinn, D. Crow, E. Manilow, and J. L. Roux, “WHAM!: Extending speech separation to noisy environments,” inINTERSPEECH, 2019, pp. 1368–1372

    2019

  28. [36]

    Continuous speech separation: Dataset and analysis,

    Z. Chen, T. Yoshioka, L. Lu, T. Zhou, Z. Meng, Y . Luo, J. Wu, X. Xiao, and J. Li, “Continuous speech separation: Dataset and analysis,” in ICASSP, 2020, pp. 7284–7288

    2020

  29. [38]

    Listen only to me! How well can target speech extraction handle false alarms?

    M. Delcroix, K. Kinoshita, T. Ochiai, K. Zmolikova, H. Sato, and T. Nakatani, “Listen only to me! How well can target speech extraction handle false alarms?” inINTERSPEECH, 2022, pp. 216–220

    2022

  30. [39]

    KenLM: Faster and smaller language model queries,

    K. Heafield, “KenLM: Faster and smaller language model queries,” in WMT, 2011, pp. 187–197

    2011

  31. [40]

    Wav2Letter++: A fast open-source speech recognition system,

    V . Pratap, A. Hannun, Q. Xu, J. Cai, J. Kahn, G. Synnaeve, V . Liptchin- sky, and R. Collobert, “Wav2Letter++: A fast open-source speech recognition system,” inICASSP, 2019, pp. 6460–6464

    2019

  32. [41]

    Improving speaker discrimination of target speech extraction with time-domain speakerbeam,

    M. Delcroix, T. Ochiai, K. Zmolikova, K. Kinoshita, N. Tawara, T. Nakatani, and S. Araki, “Improving speaker discrimination of target speech extraction with time-domain speakerbeam,” inICASSP, 2020, pp. 691–695

    2020

  33. [42]

    The proof and measurement of association between two things,

    C. Spearman, “The proof and measurement of association between two things,”The American Journal of Psychology, vol. 15, no. 1, pp. 72–101, 1904

    1904

Pith tools

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