The reviewed record of science sign in
Pith

arxiv: 2606.10233 · v1 · pith:DO6PVYKT · submitted 2026-06-08 · eess.AS · cs.LG· cs.SD

ANCHOR: Autoregressive Non-intrusive Chunk-Ordered Refinement for Joint Multi-Resolution Speech Quality Modeling

Zhuoyan Tao , Jiatong Shi , Hye-jin Shim , Shinji Watanabe This is my paper

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel 2026-06-27 14:37 UTCgrok-4.3pith:DO6PVYKTrecord.jsonopen to challenge →

classification eess.AS cs.LGcs.SD
keywords speech quality assessmentincremental predictionautoregressive modelingmulti-resolution tokensnon-intrusive evaluationpartial audio inputstreaming systems
0
0 comments X

The pith

ANCHOR predicts speech quality from partial audio prefixes by jointly modeling chunk and utterance scores in one autoregressive decoder.

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

The paper establishes that incremental speech quality assessment can be reframed as a multi-resolution autoregressive task where both chunk-level and full-utterance quality are estimated together. It does this by feeding dual-resolution tokens into a single decoder equipped with a resolution-aware hierarchy that refines coarse predictions into finer ones. A sympathetic reader would care because streaming and generative audio systems need reliable quality estimates without waiting for complete utterances, yet prior predictors degrade sharply on prefix inputs. Experiments demonstrate that this approach yields substantial robustness on short prefixes and reveals a stable perceptual context horizon after several seconds.

Core claim

ANCHOR reformulates incremental assessment as a multi-resolution autoregressive task that models chunk- and utterance-level quality within a single decoder using dual-resolution tokens and a resolution-aware hierarchy for coarse-to-fine refinement, producing reliable predictions on partial inputs where existing methods fail.

What carries the argument

dual-resolution tokens together with a resolution-aware hierarchy inside one decoder that enables coarse-to-fine refinement across chunk and utterance scales

If this is right

  • The model achieves a 48% reduction in PLCMOS error on 2-second audio prefixes compared with prior full-context predictors.
  • Perceptual quality predictions converge after an effective context horizon of 4-6 seconds.
  • Hierarchical supervision yields better incremental accuracy than single-resolution training.
  • Localized corruption produces identifiable structured extrapolation biases that can be isolated in stress tests.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same chunk-ordered refinement pattern could be tested on other incremental audio tasks such as real-time emotion or speaker verification.
  • If the 4-6 second horizon holds across languages, it would set a practical lower bound on buffer length for low-latency quality monitors.
  • The bias patterns under corruption suggest a route to add explicit uncertainty estimates when the input deviates from training distributions.

Load-bearing premise

Dual-resolution tokens plus a resolution-aware hierarchy inside a single decoder will produce reliable coarse-to-fine refinement for incremental quality prediction on unseen partial inputs.

What would settle it

Run the same PLCMOS evaluation on 2-second prefixes from a held-out corpus with different speakers or acoustic conditions and check whether the reported error reduction relative to non-hierarchical baselines disappears or reverses.

Figures

Figures reproduced from arXiv: 2606.10233 by Hye-Jin Shim, Jiatong Shi, Shinji Watanabe, Zhuoyan Tao.

Figure 1
Figure 1. Figure 1: Joint multi-resolution autoregressive modeling. Given prefix x1:t, the shared decoder generates chunk-level quality fol￾lowed by full-utterance quality. is y c (t) = mc(x1:t), while for mf ∈ Mf , the target is y f = mf (x1:T ), although the model input remains x1:t. 3.2. Resolution-Aware Decoding Order To structure the dependency between local and global quality, we impose a specific decoding order. Let Y … view at source ↗
Figure 2
Figure 2. Figure 2: Chunk-level prediction MAE across prefix lengths. (a) PLCMOS: ANCHOR consistently outperforms ARECHO. (b) UTMOS: ANCHOR wins at 2 s but a crossover emerges at longer prefixes. 4.4. Baseline Our primary baseline is the pretrained ARECHO check￾point [13], which predicts multi-metric quality tokens from com￾plete utterances using the same WavLM frontend and decoder architecture. ARECHO is trained exclusively … view at source ↗
Figure 3
Figure 3. Figure 3: Prefix-to-full convergence under ANCHOR. (a) MAE gap drops sharply between 2 s and 4 s, with diminishing returns thereafter. (b) Pearson correlation stabilizes by 4–6 s, indicating an effective context horizon [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
read the original abstract

While speech quality is typically assessed on complete utterances, streaming and generative systems require incremental estimation from partial audio. Existing predictors assume full context, degrading on prefix-constrained inputs. Extending ARECHO, we propose ANCHOR, reformulating incremental assessment as a multi-resolution autoregressive task. It models chunk- and utterance-level quality within a single decoder using dual-resolution tokens and a resolution-aware hierarchy for coarse-to-fine refinement. Experiments show substantial robustness under partial input, including a 48% PLCMOS error reduction on 2-second prefixes. Convergence analysis reveals a 4-6 s effective perceptual context horizon. A stress test further isolates structured extrapolation biases under localized corruption. Results demonstrate that hierarchical supervision improves incremental prediction and elucidates how perceptual quality accumulates over time.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper proposes ANCHOR as an extension of ARECHO for incremental, non-intrusive speech quality assessment on partial audio inputs. It reformulates the task as multi-resolution autoregressive modeling within a single decoder, employing dual-resolution tokens and a resolution-aware hierarchy to enable coarse-to-fine refinement of chunk- and utterance-level quality scores. Key reported results include a 48% reduction in PLCMOS error on 2-second prefixes, identification of a 4-6 s effective perceptual context horizon via convergence analysis, and evidence that hierarchical supervision mitigates extrapolation biases under localized corruption.

Significance. If the quantitative gains and convergence findings hold under rigorous controls, the work would meaningfully advance quality modeling for streaming and generative speech systems, where full-utterance context is unavailable. The hierarchical autoregressive formulation offers a principled way to study how perceptual quality accumulates over time, potentially informing both evaluation metrics and model design in real-time audio pipelines.

major comments (2)
  1. [Abstract / modeling description] The central modeling claim—that dual-resolution tokens plus a resolution-aware hierarchy inside one decoder reliably produces coarse-to-fine refinement for unseen partial inputs—lacks supporting ablations or architectural diagrams in the provided text; without these, it is impossible to determine whether the hierarchy is load-bearing or whether simpler concatenation of resolutions would suffice.
  2. [Experiments / results paragraph] The 48% PLCMOS error reduction on 2-second prefixes is presented without reference to the exact baseline (e.g., full-context ARECHO, non-hierarchical variants, or standard MOS predictors), data splits, or statistical significance testing; this detail is required to evaluate whether the gain is attributable to the proposed architecture rather than dataset or evaluation choices.
minor comments (1)
  1. [Abstract] The abstract is information-dense; expanding the methods paragraph with one sentence on tokenization and loss formulation would improve readability for readers unfamiliar with ARECHO.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed review and constructive comments. We address each major comment below and will revise the manuscript to improve clarity and provide the requested supporting material.

read point-by-point responses

  1. Referee: [Abstract / modeling description] The central modeling claim—that dual-resolution tokens plus a resolution-aware hierarchy inside one decoder reliably produces coarse-to-fine refinement for unseen partial inputs—lacks supporting ablations or architectural diagrams in the provided text; without these, it is impossible to determine whether the hierarchy is load-bearing or whether simpler concatenation of resolutions would suffice.

    Authors: We agree that the abstract is concise and that the provided excerpt does not include ablations or diagrams. The full manuscript describes the dual-resolution tokens and resolution-aware hierarchy, but to directly address this point we will add an architectural diagram and ablation experiments comparing the hierarchical model against a non-hierarchical baseline (simple concatenation of resolutions) to demonstrate that the hierarchy is necessary for reliable coarse-to-fine refinement on unseen partial inputs. revision: yes

  2. Referee: [Experiments / results paragraph] The 48% PLCMOS error reduction on 2-second prefixes is presented without reference to the exact baseline (e.g., full-context ARECHO, non-hierarchical variants, or standard MOS predictors), data splits, or statistical significance testing; this detail is required to evaluate whether the gain is attributable to the proposed architecture rather than dataset or evaluation choices.

    Authors: The reported 48% reduction is measured against the full-context ARECHO baseline on identical prefix inputs and the same data splits (VCTK and additional corpora described in the experiments section). We will revise the results paragraph to explicitly name all baselines (including non-hierarchical variants and standard full-utterance MOS predictors), specify the exact data splits, and add statistical significance testing (paired t-tests or bootstrap confidence intervals) to confirm the improvement is due to the ANCHOR architecture. revision: yes

Circularity Check

0 steps flagged

No circularity detected; derivation chain not present in provided text

full rationale

The supplied abstract and context contain no equations, fitted parameters, uniqueness theorems, or derivation steps that could be inspected for reduction to inputs by construction. The modeling description is high-level and refers to an extension of prior work (ARECHO) without exhibiting any self-definitional, fitted-prediction, or self-citation load-bearing structure. Absent the full manuscript's technical sections, no load-bearing step can be quoted or shown to collapse, satisfying the requirement to flag circularity only with explicit evidence. The result is therefore scored as self-contained with no circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no identifiable free parameters, axioms, or invented entities; full text would be required to audit modeling assumptions.

pith-pipeline@v0.9.1-grok · 5676 in / 970 out tokens · 17670 ms · 2026-06-27T14:37:04.025934+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references · 1 canonical work pages · 1 internal anchor

  1. [1]

    This incremental setting mirrors human speech per- ception, where listeners process acoustic signals as they unfold in time rather than waiting for utterance completion [10]

    Introduction Many practical speech systems, such as streaming communica- tion [1], speech enhancement [2], and autoregressive generative speech models [3–9], operate under partial acoustic context, mo- tivating perceptual quality estimation before full utterances are observed. This incremental setting mirrors human speech per- ception, where listeners pro...

  2. [2]

    ANCHOR: Autoregressive Non-intrusive Chunk-Ordered Refinement for Joint Multi-Resolution Speech Quality Modeling

    Related Work Traditional speech quality assessment has relied on intrusive met- rics such as PESQ [11], ViSQOL [12], and STOI [14]. While ac- curate, these methods require a clean, time-aligned reference sig- nal, limiting their utility in real-world scenarios where only the degraded signal is available. To address this, non-intrusive (no- reference) asse...

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  3. [3]

    Multi-Resolution Autoregressive Modeling ANCHOR extends ARECHO [13] by introducing dual- resolution metric query tokens and a resolution-aware decoding hierarchy

    Method and Task Formulation 3.1. Multi-Resolution Autoregressive Modeling ANCHOR extends ARECHO [13] by introducing dual- resolution metric query tokens and a resolution-aware decoding hierarchy. We leverage two properties fundamental to ARECHO: (1) a unified token space for heterogeneous perceptual metrics and (2) autoregressive conditioning that capture...

  4. [4]

    Dataset and Prefix Construction Dataset.Experiments utilize theOverall Baseconfiguration from [13], spanning 308.8 hours of clean, corrupted, and synthe- sized speech

    Experimental Setup 4.1. Dataset and Prefix Construction Dataset.Experiments utilize theOverall Baseconfiguration from [13], spanning 308.8 hours of clean, corrupted, and synthe- sized speech. The training set contains 170,013 utterances span- ning clean speech (OWSM-V3 [29 –31]), simulated/enhanced speech (URGENT2024 [32], V oiceBank+DEMAND [33, 34]), and...

    2022

  5. [5]

    Chunk-Level: Local vs

    Results We organize results around three questions: (1) Does chunk- first decoding improve prefix-based prediction? (2) How quickly does full-utterance prediction converge as context grows? (3) How do the two systems behave under controlled distortion? 5.1. Chunk-Level: Local vs. Global Trade-off Figure 2 compares ANCHOR and ARECHO on chunk-level (prefix-...

  6. [6]

    By enforcing a resolution-aware decoding hierarchy, we mitigate supervision conflict between local and global objec- tives

    Conclusion We introduced ANCHOR, a unified autoregressive framework for joint chunk-level and full-utterance speech quality model- ing. By enforcing a resolution-aware decoding hierarchy, we mitigate supervision conflict between local and global objec- tives. Experiments show that ANCHOR substantially improves packet-loss–sensitive perceptual modeling und...

  7. [7]

    National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296

    Acknowledgments This work used the Bridges-2 system at PSC and the Delta sys- tem at NCSA through allocations CIS210014 and IRI120008P from the Advanced Cyberinfrastructure Coordination Ecosys- tem: Services & Support (ACCESS) program [37 –39], which is supported by U.S. National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296

  8. [8]

    All experimental design, analysis, and scientific claims are the authors’ own, and the authors take full respon- sibility for the work and its content

    Use of Generative AI Disclosure Generative AI tools were used to assist with code implemen- tation and debugging, and for minor language editing of the manuscript. All experimental design, analysis, and scientific claims are the authors’ own, and the authors take full respon- sibility for the work and its content

  9. [9]

    Emformer: Efficient memory transformer based acoustic model for low latency streaming speech recognition,

    Y . Shi, Y . Wang, C. Wu, C.-F. Yeh, J. Chan, F. Zhang, D. Le, and M. Seltzer, “Emformer: Efficient memory transformer based acoustic model for low latency streaming speech recognition,” in ICASSP 2021 – 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 6783–6787

    2021

  10. [10]

    Real time speech enhance- ment in the waveform domain,

    A. Defossez, G. Synnaeve, and Y . Adi, “Real time speech enhance- ment in the waveform domain,” inProc. Interspeech 2020, 2020, pp. 3291–3295

    2020

  11. [11]

    Tacotron: Towards end- to-end speech synthesis,

    Y . Wang, R. J. Skerry-Ryan, D. Stanton, Y . Wu, R. J. Weiss, N. Jaitly, Z. Yang, Y . Xiao, Z. Chen, S. Bengio, Q. Le, Y . Agiomyr- giannakis, R. Clark, and R. A. Saurous, “Tacotron: Towards end- to-end speech synthesis,” inProc. Interspeech, 2017

    2017

  12. [12]

    Efficient neural audio synthesis,

    N. Kalchbrenner, E. Elsen, K. Simonyan, S. Noury, N. Casagrande, E. Lockhart, F. Stimberg, A. van den Oord, S. Dieleman, and K. Kavukcuoglu, “Efficient neural audio synthesis,” inproc. ICML, 2018

    2018

  13. [13]

    Audiolm: A language modeling approach to audio generation,

    Z. Borsos, R. Marinier, D. Vincent, E. Kharitonov, O. Pietquin, M. Sharifi, D. Roblek, O. Teboul, D. Grangier, M. Tagliasacchi, and N. Zeghidour, “Audiolm: A language modeling approach to audio generation,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 31, pp. 2523–2533, 2023

    2023

  14. [14]

    Speak, read and prompt: High-fidelity text-to-speech with minimal super- vision,

    E. Kharitonov, D. Vincent, Z. Borsos, R. Marinier, S. Girgin, O. Pietquin, M. Sharifi, M. Tagliasacchi, and N. Zeghidour, “Speak, read and prompt: High-fidelity text-to-speech with minimal super- vision,”Transactions of the Association for Computational Lin- guistics, vol. 11, pp. 1703–1718, 2023

    2023

  15. [15]

    Neural codec language models are zero-shot text to speech synthesizers,

    S. Chen, C. Wang, Y . Wu, Z. Zhang, L. Zhou, and S. Liu, “Neural codec language models are zero-shot text to speech synthesizers,” IEEE Transactions on Audio, Speech and Language Processing, vol. 33, pp. 705–718, 2025

    2025

  16. [16]

    V oicebox: Text-guided multilingual universal speech gener- ation at scale,

    M. Le, A. Vyas, B. Shi, B. Karrer, L. Sari, R. Moritz, M. Williamson, V . Manohar, Y . Adi, J. Mahadeokar, and W.-N. Hsu, “V oicebox: Text-guided multilingual universal speech gener- ation at scale,” inNeurIPS, vol. 37, 2023

    2023

  17. [17]

    Spark-tts: An efficient llm-based text-to-speech model with single-stream decoupled speech tokens,

    X. Wang, M. Jiang, Z. Ma, Z. Zhang, S. Liu, L. Li, Z. Liang, Q. Zheng, R. Wang, X. Feng, W. Bian, Z. Ye, S. Cheng, R. Yuan, Z. Zhao, X. Zhu, J. Pan, L. Xue, P. Zhu, Y . Chen, Z. Li, X. Chen, L. Xie, Y . Guo, and W. Xue, “Spark-tts: An efficient llm-based text-to-speech model with single-stream decoupled speech tokens,” arXiv Preprint, 2025

    2025

  18. [18]

    Functional parallelism in spoken word- recognition,

    W. D. Marslen-Wilson, “Functional parallelism in spoken word- recognition,”Cognition, vol. 25, no. 1-2, pp. 71–102, Mar. 1987

    1987

  19. [19]

    Per- ceptual evaluation of speech quality (PESQ)-a new method for objective intelligibility assessment of narrow-band speech signals, in the phone network,

    A. W. Rix, J. G. Beerends, M. P. Hollier, and A. P. Hekstra, “Per- ceptual evaluation of speech quality (PESQ)-a new method for objective intelligibility assessment of narrow-band speech signals, in the phone network,” inProc. ICASSP, 2001, pp. 749–752

    2001

  20. [20]

    ViSQOL: an objective speech quality model,

    A. Hines, J. Skoglund, A. C. Kokaram, and N. H. Gillis, “ViSQOL: an objective speech quality model,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 23, no. 3, pp. 524– 533, 2015

    2015

  21. [21]

    ARECHO: Autoregressive evaluation via chain-based hypothesis optimization for speech multi-metric estimation,

    J. Shi, Y . Cheng, B.-H. Su, H.-j. Shim, J. Tian, S. Cornell, Y . Zhao, S. Arora, and S. Watanabe, “ARECHO: Autoregressive evaluation via chain-based hypothesis optimization for speech multi-metric estimation,” inNeurIPS, 2025, spotlight

    2025

  22. [22]

    An al- gorithm for intelligibility prediction of time–frequency weighted noisy speech,

    C. H. Taal, R. C. Hendriks, R. Heusdens, and J. Jensen, “An al- gorithm for intelligibility prediction of time–frequency weighted noisy speech,”IEEE Transactions on Audio, Speech, and Lan- guage Processing, vol. 19, no. 7, pp. 2125–2136, 2011

    2011

  23. [23]

    Utmos: Utokyo-sarulab system for voicemos challenge 2022,

    T. Saeki, S. Takaki, and J. Yamagishi, “Utmos: Utokyo-sarulab system for voicemos challenge 2022,” inProc. Interspeech, 2022

    2022

  24. [24]

    Dnsmos: A non-intrusive perceptual objec- tive speech quality metric to evaluate noise suppressors,

    C. K. A. Reddyet al., “Dnsmos: A non-intrusive perceptual objec- tive speech quality metric to evaluate noise suppressors,” inProc. ICASSP, 2021

    2021

  25. [25]

    Mosnet: Deep learning-based objective assess- ment for voice conversion,

    C.-C. Lo, S.-W. Fu, W.-C. Huang, X. Wang, J. Yamagishi, Y . Tsao, and H.-M. Wang, “Mosnet: Deep learning-based objective assess- ment for voice conversion,” inProc. Interspeech, 2019

    2019

  26. [26]

    NORESQA: A Non-reference Speech Quality As- sessment metric using non-matching references,

    P. Manocha, A. Kumar, Z. Buyukkayhan, Z. Jin, A. Finkelstein, and G. J. Paris, “NORESQA: A Non-reference Speech Quality As- sessment metric using non-matching references,” inProc. ICASSP, 2021

    2021

  27. [27]

    Nisqa: A deep cnn-self-attention model for multidimensional speech quality pre- diction,

    G. Mittag, B. Naderi, A. Chehadi, and S. M¨oller, “Nisqa: A deep cnn-self-attention model for multidimensional speech quality pre- diction,” inProc. Interspeech, 2021

    2021

  28. [28]

    Nomad: Unsupervised learning of perceptual embeddings for speech enhancement and non-matching reference audio quality assessment,

    A. Ragano, J. Skoglund, and A. Hines, “Nomad: Unsupervised learning of perceptual embeddings for speech enhancement and non-matching reference audio quality assessment,” 2024

    2024

  29. [29]

    Songeval: A benchmark for automatic music aes- thetic evaluation,

    J. Leeet al., “Songeval: A benchmark for automatic music aes- thetic evaluation,” 2023

    2023

  30. [30]

    Meta audiobox aesthetics: Uni- fied automatic quality assessment for speech, music, and sound,

    A. Tjandra, Y .-C. Wu, B. Guo, J. Hoffman, B. Ellis, A. Vyas, B. Shi, S. Chen, M. Le, N. Zacharovet al., “Meta audiobox aesthetics: Uni- fied automatic quality assessment for speech, music, and sound,” arXiv preprint, 2025

    2025

  31. [31]

    Urgentmos: Unified multi-metric and preference learning for robust speech quality assessment,

    X. Zhanget al., “Urgentmos: Unified multi-metric and preference learning for robust speech quality assessment,”arXiv preprint, 2024

    2024

  32. [32]

    Universa: Unified and versatile evaluation for speech and audio,

    J. Shiet al., “Universa: Unified and versatile evaluation for speech and audio,” inProc. Interspeech, 2025

    2025

  33. [33]

    To- wards frame-level quality predictions of synthetic speech,

    M. Kuhlmann, F. Seebauer, P. Wagner, and R. Haeb-Umbach, “To- wards frame-level quality predictions of synthetic speech,” inProc. Interspeech, 2025

    2025

  34. [34]

    Chunk based speech pre-training with high resolution finite scalar quantization,

    Y . Tang and C. Tseng, “Chunk based speech pre-training with high resolution finite scalar quantization,” inProc. ICLR, 2025

    2025

  35. [35]

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

    A. Pasad, J.-C. Chou, and K. Livescu, “Layer-wise analysis of a self-supervised speech representation model,” inProc. ICASSP, 2021

    2021

  36. [36]

    Streaming automatic speech recognition with the transformer model,

    N. Moritz, T. Hori, and J. Le Roux, “Streaming automatic speech recognition with the transformer model,” inProc. ICASSP. IEEE, 2020, pp. 6734–6738

    2020

  37. [37]

    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,” inProc. O-COCOSDA, 2017

    2017

  38. [38]

    The ami meeting corpus,

    W. Kraaij, T. Hain, M. Lincoln, and W. Post, “The ami meeting corpus,” inProc. MLMI, 2005

    2005

  39. [39]

    Common voice: A massively-multilingual speech corpus,

    R. Ardila, M. Branson, K. Daviset al., “Common voice: A massively-multilingual speech corpus,” inProc. LREC, 2020

    2020

  40. [40]

    Urgent challenge: Uni- versality, robustness, and generalizability for speech enhancement,

    W. Zhang, R. Scheibler, K. Saijoet al., “Urgent challenge: Uni- versality, robustness, and generalizability for speech enhancement,” inProc. Interspeech, 2024

    2024

  41. [41]

    The voice bank corpus: De- sign, collection and data analysis of a large regional accent speech database,

    C. Veaux, J. Yamagishi, and S. King, “The voice bank corpus: De- sign, collection and data analysis of a large regional accent speech database,” inProc. O-COCOSDA, 2013

    2013

  42. [42]

    The diverse environments multi-channel acoustic noise database (demand),

    J. Thiemann, N. Ito, and E. Vincent, “The diverse environments multi-channel acoustic noise database (demand),” inProc. Meet- ings on Acoustics, 2013

    2013

  43. [43]

    The voicemos challenge 2022,

    W.-C. Huang, E. Cooper, Y . Tsaoet al., “The voicemos challenge 2022,” inProc. Interspeech, 2022

    2022

  44. [44]

    Wavlm: Large-scale self-supervised pre-training for full stack speech processing,

    S. Chenet al., “Wavlm: Large-scale self-supervised pre-training for full stack speech processing,”IEEE Journal of Selected Topics in Signal Processing, 2022

    2022

  45. [45]

    XSEDE: Accelerating scientific discovery,

    J. Towns, T. Cockerill, M. Dahan, I. Foster, K. Gaither, A. Grimshaw, V . Hazlewood, S. Lathrop, D. Lifka, G. D. Peterson, R. Roskies, J. R. Scott, and N. Wilkins-Diehr, “XSEDE: Accelerating scientific discovery,”Computing in Science & Engineering, vol. 16, no. 5, pp. 62–74, Sept.-Oct. 2014. [Online]. Available: https://doi.ieeecomputersociety.org/10.1109...

    2014

  46. [46]

    Bridges: a uniquely flexible HPC resource for new communities and data analytics,

    N. A. Nystrom, M. J. Levine, R. Z. Roskies, and J. R. Scott, “Bridges: a uniquely flexible HPC resource for new communities and data analytics,” inProceedings of the 2015 XSEDE Confer- ence: Scientific Advancements Enabled by Enhanced Cyberinfras- tructure, 2015, pp. 1–8

    2015

  47. [47]

    ACCESS: Advancing innovation: NSF’s advanced cyberinfrastruc- ture coordination ecosystem: Services & support,

    T. J. Boerner, S. Deems, T. R. Furlani, S. L. Knuth, and J. Towns, “ACCESS: Advancing innovation: NSF’s advanced cyberinfrastruc- ture coordination ecosystem: Services & support,” inPractice and Experience in Advanced Research Computing (PEARC ’23). New York, NY , USA: Association for Computing Machinery, 2023, pp. 1–4

    2023