Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Collective Learning Mechanism based Optimal Transport Generative Adversarial Network for Non-parallel Voice Conversion

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A single-generator GAN with three differently structured discriminators and an optimal-transport loss is claimed to outperform existing non-parallel voice-conversion models on standard benchmark corpora.

desk verdict A decent multi-discriminator GAN for voice conversion is undermined by an OT loss that does not algebraically hold together and an evaluation that is thinner than the abstract claims. read the letter →

arxiv 2504.13791 v1 pith:HCDCRRGA submitted 2025-04-18 cs.SD cs.AIeess.AS

Sandipan Dhar , Md. Tousin Akhter , Nanda Dulal Jana , Swagatam Das This is my paper · ORCID
classification cs.SDcs.AIeess.AS
keywords voiceconversiongenerativeadversarialnetworkoptimaltransportmultiplediscriminatorscollectivelearningSinkhorndivergencemel-spectrogramnon-parallel
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 aims to improve non-parallel voice conversion by replacing the usual single discriminator in a GAN with three differently structured discriminators: a deep CNN, a vision transformer, and a conformer. These discriminators are combined through a collective weighting rule, and the model also adds an optimal-transport loss that measures the gap between source and target mel-spectrogram distributions. The claim is that this combined design makes converted speech closer to the target speaker while preserving linguistic content, and the reported experiments on three standard benchmark corpora show lower Mel-Cepstral Distortion and Modulation Spectra Distance and higher Mean Opinion Scores than two prior GAN-based conversion systems. A sympathetic reader would care because the method targets a known weak point of current GAN voice converters: the naturalness gap between real and synthesized speech, and it claims to do so with a training scheme that works on small datasets. If the claim holds, multi-discriminator feedback plus distribution-level loss is a practical recipe for one-to-one non-parallel voice conversion.

What carries the argument

The load-bearing mechanism is the collective learning rule paired with the optimal-transport discriminator loss. For each discriminator, the paper computes an individual loss $L_{d_k}$ and a total $L_{\text{tot}} = \sum_k L_{d_k}$; the participation weight is $\alpha_k = (L_{\text{tot}} - L_{d_k}) / L_{\text{tot}}$, so the discriminator with the smallest loss gets the largest weight, and the final discriminator loss is $L = \sum_k \alpha_k L_{d_k}$. The optimal-transport loss is written as $L_d = W_c(X,X') + W_c(X,Y') + W_c(X',Y) + W_c(X',Y') - 2W_c(X,X') - 2W_c(Y,Y')$, where $W_c$ is the optimal transport cost with a cosine cost and $N=4$ Sinkhorn matchings. This mechanism tells each discriminator to focus on distribution mismatch rather than raw classification, and tells the generator to prioritize feedback from discriminators that are already doing well, which the paper interprets as a multiplayer cooperative game.

What would settle it

Run the exact training protocol with mini-batch size 1 and $N=4$, print the four vectors entering each $W_c$ term of Eq. (3), and compare the resulting MCD and MOS against a version trained with an $L_2$ discriminator loss under otherwise identical settings; if the improvement disappears, the claimed role of optimal transport is not supported.

Watch

Extended reading notes

Core claim

The paper's central claim is that CLOT-GAN-VC, a single-generator GAN with three discriminators updated by a collective learning mechanism and trained with an optimal-transport discriminator loss, outperforms existing one-to-one non-parallel voice-conversion models on objective and subjective metrics. The collective mechanism weights each discriminator's loss inversely to its contribution to the total loss, so the generator receives a composite feedback signal from all three architectures. The optimal-transport loss, computed from cosine transport costs and Sinkhorn soft matchings on flattened feature embeddings, is intended to align the generated mel-spectrogram distribution with the target distribution rather than simply classifying real versus fake. Ablation results are used to attribute the gains specifically to the multi-discriminator setup, the weighted averaging, and the optimal-transport loss, with the relative importance varying by gender pair.

Load-bearing premise

The paper assumes that the optimal-transport discriminator loss in Eq. (3) is correctly specified and can be computed with the stated Sinkhorn matching size $N=4$ under a mini-batch size of 1; if the formula or its implementation is inconsistent, the central explanation for the performance gain is unsupported.

Editorial extensions

If this is right

  • If the reported results hold, replacing a single discriminator with a diverse set of three architectures is a viable strategy for one-to-one non-parallel voice conversion without requiring parallel training data.
  • The ablation's ranking implies that the optimal-transport loss is the most important component for male-to-male and inter-gender conversion, while the multi-discriminator design matters most for female-to-female conversion; a practitioner could tune components by gender pair.
  • Because training used only 81 training samples per dataset, the method claims to be suitable for low-resource voice-conversion settings where large parallel corpora are unavailable.
  • The optimal-transport loss gives the generator a continuous distribution-level training signal, so the approach could reduce the over-smoothing and naturalness gap that the paper attributes to earlier GAN voice converters.

Reading between the lines

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

  • One consequence the paper does not test directly: swapping the optimal-transport discriminator loss onto a single-discriminator version would isolate whether the gain comes from the optimal-transport loss alone; the current ablations only remove optimal transport, multiple discriminators, and weighting one at a time from the full system, so the interaction between optimal transport and multiple di
  • If the collective weighting rule is sound, it should transfer to other speech-generation tasks with multiple feature extractors, such as speech enhancement or text-to-speech, where each discriminator could attend to a different acoustic property; that is an extension rather than a claim of this paper.
  • The stated mini-batch size of 1 and matching size $N=4$ leave an implementation question: where the four mini-batch vectors for the optimal-transport loss come from within a single sample. A reader checking the released code should verify this before trusting the optimal-transport explanation.
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 / 5 minor

Summary. The paper proposes CLOT-GAN-VC, a single-generator multi-discriminator GAN for non-parallel voice conversion. The generator follows the MaskCycleGAN-VC backbone, and the three discriminators (DCNN, ViT, conformer) are combined through a 'collective learning mechanism' that weights each discriminator's loss, where the discriminator loss is an optimal-transport (OT) loss based on the Sinkhorn algorithm. The authors evaluate the model on VCC 2018, CSTR-VCTK, and CMU-Arctic, reporting MCD/MSD and MOS against MaskCycleGAN-VC and MelGAN-VC, plus an ablation study on VCC 2018. The abstract claims that the model outperforms existing VC models in both objective and subjective assessments.

Significance. If the proposed method worked as described, the multi-discriminator collective-learning design and the OT-based training signal would be a reasonable and potentially useful direction for non-parallel VC, particularly the use of ViT and conformer discriminators. The paper includes an ablation study, reports MOS confidence intervals, and provides links to code and generated samples, which are useful practices. However, the mathematical specification of the central OT loss is internally inconsistent, the participation weights in the collective learning mechanism do not sum to one, and the empirical claims are overstated relative to the reported numbers. These issues are load-bearing because the novelty of the paper rests on the OT loss and the collective learning mechanism.

major comments (4)
  1. [Section III, 'Discriminator OT loss', Eq. (3)] Equation (3) is not a valid or recognizable Sinkhorn/OT divergence as written. Combining like terms reduces it to Wc(X,Y') + Wc(X',Y) + Wc(X',Y') - Wc(X,X') - 2Wc(Y,Y'), which contains a negative within-class cost for the real class and a double-negative within-class cost for the fake class. No derivation is given, and the cited references [23], [24], and [32] do not contain this expression. This matters because the ablation in Table II is used to attribute the reported gains to the OT loss, so the trained loss must be exactly what is stated. In addition, Eq. (1) sets N=4 but Section IV-A states a training mini-batch size of 1; the paper does not explain how four samples per mini-batch are formed. If the implementation in fact uses a different OT loss, Eq. (3) is a material misdescription; if it uses Eq. (3), the loss is not the claimed OT-bridging mechanism. Please correct the equation, provide a derivation or citation, and clarify the mini-batch construction, or rerun the experiments with the intended loss.
  2. [Algorithm 1 and generator loss (Eqs. (4)-(5))] The participation weights in Algorithm 1 are defined as alpha_k^y = (Ltot_y - Ld_k^y)/Ltot_y. Since Ltot_y = sum_j Ld_j^y, these weights sum to n-1, not to 1. With n=3, the weights sum to 2, so the final loss L_y = sum_k alpha_k^y Ld_k^y is not a weighted average as claimed in the text. The same issue applies to the beta weights in the generator loss. Please clarify whether normalization is intended and, if so, how it is performed; as written, the collective learning mechanism changes the loss scale and the interpretation of the weighting.
  3. [Section V-A, Table I] The abstract states that CLOT-GAN-VC outperforms existing VC models, but Table I contains several counterexamples. For example, on CMU-Arctic F-F, CLOT-GAN-VC has MCD 8.66 versus 7.81 for MaskCycleGAN-VC; on CMU-Arctic M-F, CLOT-GAN-VC has MSD 1.41 versus 1.32 for MaskCycleGAN-VC; on CSTR-VCTK F-M, CLOT-GAN-VC has MCD 5.86 versus 5.26 for MelGAN-VC; and on CSTR-VCTK M-M, CLOT-GAN-VC has MSD 1.40 versus 1.38 for MaskCycleGAN-VC. No error bars, confidence intervals, or significance tests are reported for the objective metrics, so the claim that the model outperforms the baselines is not supported as stated.
  4. [Section V-B, Table III] Table III reports MOS values with 95% confidence intervals, but many intervals overlap with those of the baselines. For instance, VCC2018 M-M gives 3.57±0.53 for CLOT-GAN-VC versus 3.47±0.50 for MaskCycleGAN-VC, and CMU-Arctic M-M gives 3.22±0.09 versus 3.16±0.03. The text states that the proposed model is 'significantly higher' without any statistical significance test. With only 17 raters and overlapping intervals, the subjective superiority claim is not established. Please perform an appropriate significance test or temper the claim.
minor comments (5)
  1. [Section IV-A] The sentence 'The training, validation and evaluation (test) sets for each dataset comprised of 81, 35, 25 samples' is unclear: it does not specify whether these counts are per speaker pair, per utterance, or per dataset, and 81/35/25 is an unusual split. Please clarify.
  2. [References] References [12] and [13] are duplicates (same title, authors, and conference). Please remove one.
  3. [Section V-A, ablation discussion] The rank-based interpretation of Table II is informal and no error bars are given for the ablation runs, so the stated conclusions about which component contributes most are not statistically grounded.
  4. [Figure 2] The mel-spectrogram and Grad-CAM panels in Figure 2 are extremely small, making it difficult to verify the claimed visual differences. Please enlarge the figure or provide higher-resolution views.
  5. [Reproducibility links] The code and demo links use short URLs (shorturl.at, tinyurl.com), which are not stable for archival purposes. Please provide persistent repository links or DOIs.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is an empirical method proposal whose claims are tested against external baselines, and no derivation reduces to its own inputs.

full rationale

This paper is an empirical method paper. The central claim is that the proposed CLOT-GAN-VC model outperforms baseline VC models on VCC 2018, VCTK, and CMU-Arctic; that claim is supported by objective and subjective measurements against external baselines (MaskCycleGAN-VC and MelGAN-VC) and by ablations, not by a derivation from assumptions that include the target conclusion. The OT loss is introduced as a definition in Eqs. (1)-(3), not derived from the experimental outcome, and the collective weighting in Algorithm 1 is a constructive training rule rather than a prediction. The cycle and identity losses are said to be 'kept similar to FID-RPRGAN-VC [16]', which is a self-citation, but it is not load-bearing: those are standard auxiliary losses, and the paper's central mechanism (multi-discriminator weighting, OT loss, and ablation evidence) does not rest on [16] as a theorem. The apparent sign inconsistency in Eq. (3) and the stated N=4 vs. mini-batch-size-1 mismatch are implementation and correctness concerns, not circularity: they do not make an output equal to an input by construction. There is no fitted parameter renamed as a prediction, no imported uniqueness theorem, and no instance where a stated result is equivalent to its own premise by definition. A non-finding is therefore appropriate.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities. The collective learning mechanism is an algorithmic weighting rule, not a postulated entity. The load-bearing assumptions are the correctness of the OT loss definition, the validity of the Sinkhorn approximation, the transferability of cycle losses from prior work, the reliability of the vocoder, and the choice of cosine ground cost.

free parameters (3)
  • number of discriminators n = 3
    Chosen by hand; the paper does not ablate the number of discriminators, only whether multiple discriminators are used at all.
  • OT cost matrix size N = 4
    Stated as N=4 for the Sinkhorn matchings in Eq. (1), but training uses mini-batch size 1; the inconsistency is unexplained and the value is not justified.
  • training learning rate = 1e-4
    Standard Adam learning rate used in training; it is a hyperparameter, not the focus of the claim, but part of the experimental setup.
assumptions (5)
  • ad hoc to paper The OT loss expression in Eq. (3) correctly implements an optimal transport divergence.
    The expression simplifies to -Wc(X,X') + Wc(X,Y') + Wc(X',Y) + Wc(X',Y') - 2Wc(Y,Y') and is not a standard OT discriminator loss; no derivation is provided.
  • standard math The Sinkhorn algorithm converges to a valid soft matching matrix M for the cost matrix C.
    Sinkhorn scaling is a standard method for entropic OT approximation; the paper relies on it to compute Wc.
  • domain assumption Cycle consistency and identity losses from FID-RPRGAN-VC transfer to the proposed model.
    Section III states these losses are kept similar to [16], the authors' prior work; their suitability is assumed without re-derivation.
  • domain assumption The pretrained MelGAN vocoder reconstructs intelligible speech from generated mel-spectrograms.
    The paper uses a pretrained MelGAN vocoder for speech reconstruction and assumes it does not introduce a significant bottleneck or bias.
  • domain assumption Cosine distance is a valid ground cost c for mel-spectrogram feature embeddings.
    The cost function is chosen as cosine distance without experimental justification for this particular feature space.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Collective Learning Mechanism based Optimal Transport Generative Adversarial Network for Non-parallel Voice Conversion." pith.science (2026). https://pith.science/paper/HCDCRRGA

@misc{pith2026250413791,
  author       = {Pith},
  title        = {Pith review of: Collective Learning Mechanism based Optimal Transport Generative Adversarial Network for Non-parallel Voice Conversion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HCDCRRGA}},
  note         = {Machine review of arXiv:2504.13791}
}
read the original abstract

After demonstrating significant success in image synthesis, Generative Adversarial Network (GAN) models have likewise made significant progress in the field of speech synthesis, leveraging their capacity to adapt the precise distribution of target data through adversarial learning processes. Notably, in the realm of State-Of-The-Art (SOTA) GAN-based Voice Conversion (VC) models, there exists a substantial disparity in naturalness between real and GAN-generated speech samples. Furthermore, while many GAN models currently operate on a single generator discriminator learning approach, optimizing target data distribution is more effectively achievable through a single generator multi-discriminator learning scheme. Hence, this study introduces a novel GAN model named Collective Learning Mechanism-based Optimal Transport GAN (CLOT-GAN) model, incorporating multiple discriminators, including the Deep Convolutional Neural Network (DCNN) model, Vision Transformer (ViT), and conformer. The objective of integrating various discriminators lies in their ability to comprehend the formant distribution of mel-spectrograms, facilitated by a collective learning mechanism. Simultaneously, the inclusion of Optimal Transport (OT) loss aims to precisely bridge the gap between the source and target data distribution, employing the principles of OT theory. The experimental validation on VCC 2018, VCTK, and CMU-Arctic datasets confirms that the CLOT-GAN-VC model outperforms existing VC models in objective and subjective assessments.

Figures

Figures reproduced from arXiv: 2504.13791 by the authors.

Figure 1
Figure 1. Overview of the proposed CLOT-GAN-VC model [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Mel-spectrograms (M to F VC), generated using CLOT-GAN-VC, MaskCycleGAN-VC, and MelGAN-VC, along with Grad-CAM-based visualization [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

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. Generative Adversarial Network based Voice Conversion: Techniques, Challenges, and Recent Advancements

    cs.SD 2025-04 conditional novelty 1.0 of 10

    A narrative review of GAN-based voice conversion that maps the existing landscape but leans heavily on the authors' own models for its 'recent advancements' section.

Reference graph

Works this paper leans on

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

  1. [23]

    Towards solving nlp tasks with optimal transport loss,

    R. Bhardwaj, T. Vaidya, and S. Poria, “Towards solving nlp tasks with optimal transport loss,” Journal of King Saud University - Computer and Information Sciences, vol. 34, no. 10, Part B, pp. 10 434–10 443, 2022

    work page 2022

  2. [24]

    Hilbert sinkhorn divergence for optimal transport,

    Q. Li, Z. Wang, G. Li, J. Pang, and G. Xu, “Hilbert sinkhorn divergence for optimal transport,” in Proc. 2021 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR) , 2021, pp. 3834–3843

    work page 2021

  3. [32]

    Improving gans using optimal transport,

    T. Salimans, H. Zhang, A. Radford, and D. N. Metaxas, “Improving gans using optimal transport,” ArXiv, vol. abs/1803.05573, 2018

    arXiv 2018

  4. [1]

    An overview of voice con- version and its challenges: From statistical modeling to deep learning,

    B. Sisman, J. Yamagishi, S. King, and H. Li, “An overview of voice con- version and its challenges: From statistical modeling to deep learning,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 29, pp. 132–157, 2021

    work page 2021

  5. [2]

    V oice conversion for whispered speech synthesis,

    M. Cotescu, T. Drugman, G. Huybrechts, J. Lorenzo-Trueba, and A. Moinet, “V oice conversion for whispered speech synthesis,” IEEE Signal Processing Letters , vol. 27, pp. 186–190, 2020

    work page 2020

  6. [3]

    V oice conversion using gaussian mixture models,

    K. D’souza and K. Talele, “V oice conversion using gaussian mixture models,” in Proc. 2015 International Conference on Communication, Information and Computing Technology (ICCICT) , 2015, pp. 1–5

    work page 2015

  7. [4]

    Cross-lingual voice conversion with bilingual phonetic posteriorgram and average modeling,

    Y . Zhou, X. Tian, H. Xu, R. K. Das, and H. Li, “Cross-lingual voice conversion with bilingual phonetic posteriorgram and average modeling,” in Proc. 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2019, pp. 6790–6794

    work page 2019

  8. [5]

    Restricted boltzmann machine-based voice conversion for nonparallel corpus,

    K.-S. Lee, “Restricted boltzmann machine-based voice conversion for nonparallel corpus,” IEEE Signal Processing Letters , vol. 24, no. 8, pp. 1103–1107, 2017

    work page 2017

Show all 42 references
  1. [6]

    Mixture of factor analyzers using priors from non-parallel speech for voice conversion,

    Z. Wu, T. Kinnunen, E. S. Chng, and H. Li, “Mixture of factor analyzers using priors from non-parallel speech for voice conversion,”IEEE Signal Processing Letters, vol. 19, no. 12, pp. 914–917, 2012

    2012

  2. [7]

    Phoneme independent hmm voice conversion,

    W. Percybrooks, E. Moore, and C. McMillan, “Phoneme independent hmm voice conversion,” in Proc. 2013 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2013, pp. 6925– 6929

    2013

  3. [8]

    V oice conversion through vector quantization,

    M. Abe, S. Nakamura, K. Shikano, and H. Kuwabara, “V oice conversion through vector quantization,” in Proc. 1988 IEEE International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP) , 1988, pp. 655–658 vol.1

    1988

  4. [9]

    Generative adversarial nets,

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial nets,” in Advances in Neural Information Processing Systems 27 , 2014, pp. 2672–2680

    2014

  5. [10]

    An adaptive-learning-based generative adversarial network for one-to-one voice conversion,

    S. Dhar, N. D. Jana, and S. Das, “An adaptive-learning-based generative adversarial network for one-to-one voice conversion,”IEEE Transactions on Artificial Intelligence , vol. 4, no. 1, pp. 92–106, 2023

    2023

  6. [11]

    Parallel-data-free voice conversion using cycle-consistent adversarial networks,

    T. Kaneko and H. Kameoka, “Parallel-data-free voice conversion using cycle-consistent adversarial networks,” ArXiv, vol. abs/1711.11293,

    arXiv

  7. [12]

    V ocoder-free end-to-end voice conversion with transformer network,

    J.-W. Kim, H.-Y . Jung, and M. Lee, “V ocoder-free end-to-end voice conversion with transformer network,” in 2020 International Joint Con- ference on Neural Networks (IJCNN) , 2020, pp. 1–8

    2020

  8. [13]

    V ocoder-free end-to-end voice conversion with transformer net- work,

    ——, “V ocoder-free end-to-end voice conversion with transformer net- work,” in 2020 International Joint Conference on Neural Networks (IJCNN), 2020, pp. 1–8

    2020

  9. [14]

    Maskcyclegan-VC: Learning non-parallel voice conversion with filling in frames,

    T. Kaneko, H. Kameoka, K. Tanaka, and N. Hojo, “Maskcyclegan-VC: Learning non-parallel voice conversion with filling in frames,” Proc. 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5919–5923, 2021

    2021

  10. [15]

    V oice conversion using feature specific loss function based self-attentive generative adversarial network,

    S. Dhar, P. Banerjee, N. D. Jana, and S. Das, “V oice conversion using feature specific loss function based self-attentive generative adversarial network,” in Proc. 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  11. [16]

    FID- RPRGAN-VC: Fr ´echet inception distance loss based region-wise posi- tion normalized relativistic gan for non-parallel voice conversion,

    S. Dhar, M. T. Akhter, P. Banerjee, N. D. Jana, and S. Das, “FID- RPRGAN-VC: Fr ´echet inception distance loss based region-wise posi- tion normalized relativistic gan for non-parallel voice conversion,” in Proc. 2023 Asia Pacific Signal and Information Processing Association ...

    2023

  12. [17]

    GLGAN-VC: A guided loss based generative adversarial network for many-to-many voice conversion,

    S. Dhar, N. D. Jana, and S. Das, “GLGAN-VC: A guided loss based generative adversarial network for many-to-many voice conversion,” IEEE Transactions on Neural Networks and Learning Systems, Accepted, 2023, DOI 10.1109/TNNLS.2023.3335119

    2023

  13. [18]

    Study of speech recognition using cnn,

    R. B. Pittala, B. Tejopriya, and E. Pala, “Study of speech recognition using cnn,” in Proc. 2022 Second International Conference on Artificial Intelligence and Smart Energy (ICAIS) , 2022, pp. 150–155

    2022

  14. [19]

    ViT-GAN: Using vision transformer as discriminator with adaptive data augmentation,

    S. Hirose, N. Wada, J. Katto, and H. Sun, “ViT-GAN: Using vision transformer as discriminator with adaptive data augmentation,” in Proc. 2021 3rd International Conference on Computer Communication and the Internet (ICCCI) , 2021, pp. 185–189

    2021

  15. [20]

    Convolutional feature based vision transformer model for speech command recognition,

    S. Das, S. Dhar, and N. D. Jana, “Convolutional feature based vision transformer model for speech command recognition,” in 2023 IEEE 20th India Council International Conference (INDICON), 2023, pp. 228–232

    2023

  16. [21]

    CMGAN: Conformer-based metric gan for speech enhancement,

    R. Cao, S. Abdulatif, and B. Yang, “CMGAN: Conformer-based metric gan for speech enhancement,” in Proc. Interspeech, 2022

    2022

  17. [22]

    Nips 2016 tutorial: Generative adversarial networks,

    I. J. Goodfellow, “Nips 2016 tutorial: Generative adversarial networks,” ArXiv, vol. abs/1701.00160, 2016

    2016 arXiv

  18. [25]

    Learning with a wasserstein loss,

    C. Frogner, C. Zhang, H. Mobahi, M. Araya-Polo, and T. A. Poggio, “Learning with a wasserstein loss,” ArXiv, vol. abs/1506.05439, 2015

    2015 arXiv

  19. [26]

    Multi-agent diverse generative adversarial networks,

    A. Ghosh, V . Kulharia, V . P. Namboodiri, P. H. Torr, and P. K. Dokania, “Multi-agent diverse generative adversarial networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2018

    2018

  20. [27]

    Evolutionary game- theoretical analysis for general multiplayer asymmetric games,

    X. Zhang, P. Peng, Y . Zhou, H. Wang, and W. Li, “Evolutionary game- theoretical analysis for general multiplayer asymmetric games,” ArXiv, vol. abs/2206.11114, 2022

    2022 arXiv

  21. [28]

    MelGAN-VC: V oice conversion and audio style trans- fer on arbitrarily long samples using spectrograms,

    M. Pasini, “MelGAN-VC: V oice conversion and audio style trans- fer on arbitrarily long samples using spectrograms,” ArXiv, vol. abs/1910.03713, 2019

    1910 arXiv

  22. [29]

    Cyclegean: Cycle generative enhanced adversarial network for voice conversion,

    X. Zhang, J. Wang, N. Cheng, E. Xiao, and J. Xiao, “Cyclegean: Cycle generative enhanced adversarial network for voice conversion,” in Proc. 2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), 2021, pp. 930–937

    2021

  23. [30]

    Cyclegan-VC2: Improved cyclegan-based non-parallel voice conversion,

    T. Kaneko, H. Kameoka, K. Tanaka, and N. Hojo, “Cyclegan-VC2: Improved cyclegan-based non-parallel voice conversion,” in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2019, pp. 6820–6824

    2019

  24. [31]

    Image fine-grained inpainting,

    Z. Hui, J. Li, X. Wang, and X. Gao, “Image fine-grained inpainting,” ArXiv, vol. abs/2002.02609, 2020

    2002 arXiv

  25. [33]

    Flattening layer pruning in convo- lutional neural networks,

    E. Jeczmionek and P. A. Kowalski, “Flattening layer pruning in convo- lutional neural networks,” Symmetry, vol. 13, no. 7, 2021

    2021

  26. [34]

    The voice conversion challenge 2018: Promoting development of parallel and nonparallel methods,

    J. Lorenzo-Trueba, J. Yamagishi, T. Toda, D. Saito, F. Villavicencio, T. H. Kinnunen, and Z. Ling, “The voice conversion challenge 2018: Promoting development of parallel and nonparallel methods,”ArXiv, vol. abs/1804.04262, 2018

    2018 arXiv

  27. [35]

    Cstr vctk corpus: English multi-speaker corpus for cstr voice cloning toolkit (version 0.92),

    J. Yamagishi, C. Veaux, and K. MacDonald, “Cstr vctk corpus: English multi-speaker corpus for cstr voice cloning toolkit (version 0.92),” Sound dataset, 2019. [Online]. Available: https://doi.org/10.7488/ds/2645

    2019 doi

  28. [36]

    The cmu arctic speech databases,

    J. Kominek and A. W. Black, “The cmu arctic speech databases,” in Proc. Speech Synthesis Workshop , 2004

    2004

  29. [37]

    Non-parallel voice conversion based on free-energy minimization of speaker-conditional restricted boltzmann machine,

    T. Kishida and T. Nakashika, “Non-parallel voice conversion based on free-energy minimization of speaker-conditional restricted boltzmann machine,” in Proc. 2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) , 2022, pp. 251–255

    2022

  30. [38]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” CoRR, vol. abs/1412.6980, 2014

    2014 arXiv

  31. [39]

    MelGAN: Generative adversarial networks for conditional waveform synthesis,

    K. Kumar, R. Kumar, T. de Boissi `ere, L. Gestin, W. Z. Teoh, J. M. R. Sotelo, A. de Br ´ebisson, Y . Bengio, and A. C. Courville, “MelGAN: Generative adversarial networks for conditional waveform synthesis,” in NeurIPS, 2019

    2019

  32. [40]

    Grad-cam: Visual explanations from deep networks via gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-cam: Visual explanations from deep networks via gradient-based localization,” International Journal of Computer Vision , vol. 128, no. 2, pp. 336–359, February 2020

    2020

  33. [41]

    Nonparallel voice conversion based on phoneme classification and eigenvoices,

    Y .-p. Li, Z. Ling-hua, and D. Hui, “Nonparallel voice conversion based on phoneme classification and eigenvoices,” in Proc. 2010 IEEE 12th International Conference on Communication Technology, 2010, pp. 662– 665

    2010

  34. [2017]

    Available: https://api.semanticscholar.org/CorpusID: 26022031

    [Online]. Available: https://api.semanticscholar.org/CorpusID: 26022031

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.