Pith. sign in

REVIEW 3 major objections 4 minor 34 references

FADEL: Uncertainty-aware Fake Audio Detection with Evidential Deep Learning

T0 review · 3 major / 4 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 This paper claims that replacing the softmax classification head with Dirichlet-evidence training makes fake-audio detectors more accurate on unseen spoofing attacks and gives them uncertainty scores that track attack difficulty.

desk verdict A legitimate first application of evidential learning to anti-spoofing, but the headline gain over AASIST may rest on a weak baseline (1.47% vs. the published 1.06%), so the experimental evidence as presented doesn't yet establish the claim. read the letter →

arxiv 2504.15663 v1 pith:L6Q2RITL submitted 2025-04-22 eess.AS cs.AI

Ju Yeon Kang , Ji Won Yoon , Semin Kim , Min Hyun Han , Nam Soo Kim This is my paper · ORCID
classification eess.AScs.AI
keywords fakeaudiodetectionanti-spoofingevidentialdeeplearninguncertaintyestimationout-of-distributiongeneralizationDirichletdistributionASVspoof
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

Fake audio detectors must catch spoofing attacks that were never seen in training, but softmax classifiers tend to be overconfident on such out-of-distribution inputs. This paper proposes FADEL, a training scheme that replaces the softmax classification head with evidential deep learning: the backbone outputs non-negative evidence, which defines a Dirichlet distribution over the bonafide and spoofed class probabilities, and predictions are made from that distribution's mean while uncertainty is read from its shape. Applied to two existing detectors, the scheme lowers equal error rates on ASVspoof2019 LA (AASIST from 1.47% to 1.21% EER; Res-TSSDNet from 3.53% to 2.92%) and on cross-dataset ASVspoof2021 LA (AASIST from 8.08% to 5.60%, beating a 6.10% sharpness-aware baseline). The paper also shows that the average uncertainty across spoofing algorithms tracks the error rate, evidence that the uncertainty estimate is meaningful. If the claim holds, detector calibration improves without changing backbone architecture or inference cost.

What carries the argument

The load-bearing object is the Dirichlet distribution over class probabilities, used as a subjective-logic prior. For each input, the backbone's last-layer logits pass through a non-negative activation to become evidence $e_k$; the Dirichlet concentration parameters are $\alpha_k = e_k + 1$. The expected class probability is $\bar{p}_k = \alpha_k / S_i$ with $S_i = \sum_j \alpha_j$, and uncertainty is $u = K / S_i$, so zero-evidence inputs yield maximal uncertainty. The training loss is the weighted Bayes risk of cross-entropy under this prior, $\sum_j w_j y_{ij}(\psi(S_i) - \psi(\alpha_{ij}))$, which encourages high evidence for the correct class and low evidence otherwise. This machinery transfers onto any backbone by replacing the softmax head and loss, which is why the paper can report gains on two different detectors.

What would settle it

Re-train AASIST and AASIST-FADEL with an epoch and learning-rate sweep at matched budgets (e.g., 100, 200, and 400 epochs); the uncertainty explanation is only required if the gap persists at every matched budget, or if the evidential loss with a tuned schedule still beats the best tuned weighted-cross-entropy run.

Watch

Extended reading notes

Core claim

The central claim is that the overconfidence of maximum-likelihood softmax training is a measurable cause of fake-audio detection failures in out-of-distribution spoofing scenarios, and that evidential training removes it. In FADEL, the model's final logits are converted into non-negative evidence $e_k$ by an activation like softplus or exponential; the Dirichlet parameters are $\alpha_k = e_k + 1$, class probabilities are the Dirichlet means $\bar{p}_k = \alpha_k / \sum_j \alpha_j$, and uncertainty is $u = K / \sum_j \alpha_j$. The training objective is the Bayes risk of weighted cross-entropy under the Dirichlet prior, which reduces to $\sum_j w_j y_{ij}(\psi(S_i) - \psi(\alpha_{ij}))$. With identical architectures and hyperparameters, AASIST-FADEL reaches 1.21% EER on ASVspoof2019 LA (baseline AASIST: 1.47%) and 5.60% EER on cross-dataset ASVspoof2021 LA (AASIST: 8.08%; ASAM: 6.10%). The paper takes the resulting flattening of predicted probabilities away from 0 and 1, together with the correlation between average uncertainty and per-algorithm EER, as evidence that the model has become both more accurate and better aware of when it is unsure.

Load-bearing premise

The paper assumes the original backbone training schedule (100 epochs, optimizer settings, and data configuration) is still right after the loss is changed from weighted cross-entropy to evidential loss, so the gains measure the loss change rather than a mismatch in training setup.

Editorial extensions

If this is right

  • Across in-domain evaluation, FADEL improves AASIST's EER from 1.47% to 1.21% and Res-TSSDNet's EER from 3.53% to 2.92% on ASVspoof2019 LA.
  • In cross-dataset evaluation from ASVspoof2019 LA to ASVspoof2021 LA, AASIST-FADEL reaches 5.60% EER, beating AASIST (8.08%) and ASAM (6.10%).
  • Predicted probabilities for spoof samples spread away from the extreme bins near 0 and 1, while bonafide samples retain higher probabilities, reducing misclassifications on hard examples.
  • Average uncertainty across spoofing algorithms A07-A19 correlates with per-algorithm EER, meaning the uncertainty value can serve as a per-attack difficulty signal.
  • Choice of evidence activation matters only slightly: exponential gives the best EER (1.16%), softplus the best min t-DCF (0.0276) in the ablation.

Reading between the lines

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

  • Because FADEL is a drop-in replacement for the loss and output head, the same evidence mechanism could be applied to other anti-spoofing backbones (e.g., RawNet2, transformer-based detectors) and to deepfake audio benchmarks beyond ASVspoof; the paper does not test this.
  • The uncertainty value could plausibly be used as an operational rejection threshold in ASV pipelines, flagging high-uncertainty utterances for human review rather than binary accept/reject; the paper only demonstrates correlation, not this downstream use.
  • The reported correlation between uncertainty and EER might also be exploitable for few-shot attack triage during model development, but the paper does not establish that uncertainty is causally related to errors beyond the correlation shown.
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

3 major / 4 minor

Summary. The paper proposes FADEL, a training scheme for fake audio detection that replaces softmax-based weighted cross-entropy with an evidential deep learning objective. The model outputs are treated as evidence defining a Dirichlet distribution over class probabilities, and the loss is the expected weighted cross-entropy under this distribution (Eqs. 6-7). The method is applied to two backbones, Res-TSSDNet and AASIST, and evaluated on ASVspoof2019 LA and cross-dataset ASVspoof2021 LA. The authors report improved EER and min t-DCF over the baselines, reduced overconfidence in predicted probabilities, and a qualitative correlation between average uncertainty and per-algorithm EER.

Significance. If the empirical results hold, FADEL would be a simple, architecture-agnostic plug-in loss that mitigates softmax overconfidence and improves out-of-distribution generalization in a security-sensitive task. The mathematical formulation in Eqs. (6)-(7) is a standard weighted expected cross-entropy under a Dirichlet prior and appears correct. However, the contribution is primarily an application of existing evidential deep learning to audio anti-spoofing rather than a new derivation, and the empirical case is weakened by a likely baseline discrepancy, missing variance reporting, and an unquantified correlation claim. The paper does not provide code or checkpoints, so the reproducibility of the reported gains cannot currently be verified.

major comments (3)
  1. [Table I / Section III-B] The reported AASIST baseline (1.47% EER avg, 1.35% best) is substantially worse than the 1.06% EER reported in the original AASIST paper [20] on the same ASVspoof2019 LA evaluation protocol. Since Section III-B states that FADEL uses the original architecture and hyperparameter configurations of the backbone, the discrepancy is not explained by the described setup. If the published AASIST figure is the correct comparison point, then AASIST-FADEL's 1.21% EER is a regression rather than an improvement, which would invalidate the central claim in the abstract. The authors should reconcile this gap, report per-seed results for the baseline and FADEL, and release code and checkpoints so the comparison can be audited.
  2. [Section III-B / Tables I-III] The manuscript states that all experiments were run three times with different seeds, but only average and best results are reported, with no standard deviations, confidence intervals, or significance tests. Without these, the claim in Section IV-A that FADEL 'significantly improves' performance cannot be assessed, especially when the reported differences are small, as in Table III where best EER values are 1.13%, 1.13%, and 1.18% across activation functions. Please include variance estimates and, where appropriate, paired significance tests.
  3. [Section IV-C / Fig. 3] The claimed 'strong correlation' between average uncertainty and EER is not quantified: no correlation coefficient, confidence interval, or statistical test is reported, and the text explicitly excludes A07, A16, and A17 as exceptions. Excluding three of thirteen points without a principled criterion makes the correlation claim unverifiable. Report the correlation (e.g., Pearson and/or Spearman) over all spoofing algorithms A07-A19, and justify any exclusion of data points.
minor comments (4)
  1. [Section II-D / Eq. (6)] In Eq. (6), B is referred to as the beta function, but for a K-class Dirichlet prior it should be the multivariate beta function; please clarify.
  2. [Section III-B] The class weight ratio is stated as 1:9 for spoof to bonafide, but the exact scalar weights wj are not defined; please specify the values used in Eq. (2) and Eq. (7).
  3. [Table II] The AASIST baseline row in the cross-dataset table is not explicitly described as being retrained under the same protocol (including class weighting and hyperparameters) as AASIST-FADEL; please state the training setup for all rows.
  4. [Fig. 2] The histograms would be easier to interpret if the caption stated which subset (evaluation or development) the probabilities are computed over and what bin width is used.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: FADEL is a direct application of an external EDL objective; its uncertainty measure is defined, not fitted, and its empirical claims rest on benchmark comparisons rather than on a self-referential derivation.

full rationale

FADEL imports the evidential deep learning machinery from the external reference [30] (Sensoy et al.) and applies it to the AASIST and Res-TSSDNet backbones. The class probabilities in Eq. (5) and the uncertainty in Eq. (4) are both explicit functions of the Dirichlet parameters α, so the claim that predictions 'reflect' uncertainty is a design definition, not a derived theorem that assumes its own conclusion. The training objective in Eqs. (6)-(7) is the expected weighted cross-entropy under the Dirichlet prior; no parameter is fitted to the reported EER values, and no constant is tuned to force the uncertainty-EER correlation. That correlation is an empirical post-hoc observation across spoofing algorithms A07-A19, and the paper admits exceptions (A07, A16, A17), so it is not imposed by construction. The comparison with AASIST and ASAM uses external baselines and there is no self-citation chain: none of the cited backbone or EDL references is authored by the present authors. The skeptic's concern that the AASIST baseline (1.47% EER) is weaker than the published 1.06% EER is a legitimate experimental-control issue about whether the comparison is apples-to-apples, but it does not fall into any circularity pattern: a mismatched baseline is not an equation reducing to its own inputs. Therefore the paper is not circular; the main risks are reproducibility and baseline fidelity, not self-referential derivation.

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

The central contribution rests on applying a prior EDL framework; no new physical or conceptual entities are introduced. The free parameters are the class weight ratio and the activation function, both chosen by hand or by ablation rather than derived from theory.

free parameters (2)
  • class weight ratio w_spoof:w_bonafide = 1:9
    Chosen by hand to handle class imbalance and used in both the baseline WCE loss and the FADEL evidential loss (Section III-B).
  • evidence activation function = softplus for main results; exponential and ReLU tested in ablation
    The choice of evidence activation affects EER and min t-DCF (Table III), and the main results use softplus despite exponential giving lower EER, so the selection is not fully justified.
assumptions (3)
  • domain assumption Weighted expected cross-entropy under a Dirichlet prior (Eq. 6-7) yields meaningful evidence estimates for OOD detection.
    Taken as a given from Sensoy et al. [30]; the paper does not re-derive or validate the EDL training objective in the audio domain.
  • domain assumption The ASVspoof2019 LA evaluation set's unseen attacks (A07-A19) represent a valid out-of-distribution scenario.
    Standard benchmark assumption used to claim OOD generalization (Section III-A).
  • domain assumption The Dirichlet mean p_bonafide can be thresholded directly for EER computation, without recalibration.
    FADEL replaces softmax probabilities with p_bonafide for decision-making, but does not demonstrate that these scores are calibrated for threshold-based metrics (Sections II-C and III-A).

how reviews work

0 comments
Cite this review

Pith. "Pith review of FADEL: Uncertainty-aware Fake Audio Detection with Evidential Deep Learning." pith.science (2026). https://pith.science/paper/L6Q2RITL

@misc{pith2026250415663,
  author       = {Pith},
  title        = {Pith review of: FADEL: Uncertainty-aware Fake Audio Detection with Evidential Deep Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6Q2RITL}},
  note         = {Machine review of arXiv:2504.15663}
}
read the original abstract

Recently, fake audio detection has gained significant attention, as advancements in speech synthesis and voice conversion have increased the vulnerability of automatic speaker verification (ASV) systems to spoofing attacks. A key challenge in this task is generalizing models to detect unseen, out-of-distribution (OOD) attacks. Although existing approaches have shown promising results, they inherently suffer from overconfidence issues due to the usage of softmax for classification, which can produce unreliable predictions when encountering unpredictable spoofing attempts. To deal with this limitation, we propose a novel framework called fake audio detection with evidential learning (FADEL). By modeling class probabilities with a Dirichlet distribution, FADEL incorporates model uncertainty into its predictions, thereby leading to more robust performance in OOD scenarios. Experimental results on the ASVspoof2019 Logical Access (LA) and ASVspoof2021 LA datasets indicate that the proposed method significantly improves the performance of baseline models. Furthermore, we demonstrate the validity of uncertainty estimation by analyzing a strong correlation between average uncertainty and equal error rate (EER) across different spoofing algorithms.

Figures

Figures reproduced from arXiv: 2504.15663 by the authors.

Figure 1
Figure 1. An Overview of FADEL. more robust predictions when facing OOD spoofing attacks. Experimental results show that FADEL significantly improves the performance of fake audio detection baselines on the ASVspoof2019 Logical Access (LA) dataset. In cross-dataset evaluation, where the model is trained on ASVspoof2019 LA and tested on ASVspoof2021 LA, FADEL also outperforms other approaches. Additionally, we demonstrate the … view at source ↗
Figure 2
Figure 2. Histograms of probabilities for the bonafide class, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Scatter plot showing the relationship between aver [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

34 extracted references · 9 canonical work pages

  1. [20]

    Aasist: Audio anti-spoofing using integrated spectro-temporal graph attention networks,

    J.-w. Jung, H.-S. Heo, H. Tak, H.-j. Shim, J. S. Chung, B.-J. Lee, H.-J. Yu, and N. Evans, “Aasist: Audio anti-spoofing using integrated spectro-temporal graph attention networks,” in ICASSP 2022-2022 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, 2022, pp. 6367–6371

    2022

  2. [1]

    Natural tts synthesis by conditioning wavenet on mel spectrogram predictions,

    J. Shen, R. Pang, R. J. Weiss, M. Schuster, N. Jaitly, Z. Yang, Z. Chen, Y . Zhang, Y . Wang, R. Skerrv-Ryan et al. , “Natural tts synthesis by conditioning wavenet on mel spectrogram predictions,” in 2018 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, 2018, pp. 4779–4783

    2018

  3. [2]

    Fastspeech 2: Fast and high-quality end-to-end text to speech,

    Y . Ren, C. Hu, X. Tan, T. Qin, S. Zhao, Z. Zhao, and T.-Y . Liu, “Fastspeech 2: Fast and high-quality end-to-end text to speech,” arXiv preprint arXiv:2006.04558, 2020

    arXiv 2006

  4. [3]

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

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

    arXiv 2018

  5. [4]

    V oice conversion challenge 2020: Intra- lingual semi-parallel and cross-lingual voice conversion,

    Y . Zhao, W.-C. Huang, X. Tian, J. Yamagishi, R. K. Das, T. Kin- nunen, Z. Ling, and T. Toda, “V oice conversion challenge 2020: Intra- lingual semi-parallel and cross-lingual voice conversion,” arXiv preprint arXiv:2008.12527, 2020

    arXiv 2020

  6. [5]

    Advances in anti-spoofing: from the perspective of asvspoof challenges,

    M. R. Kamble, H. B. Sailor, H. A. Patil, and H. Li, “Advances in anti-spoofing: from the perspective of asvspoof challenges,” APSIPA Transactions on Signal and Information Processing , vol. 9, p. e2, 2020

    2020

  7. [6]

    On the vulnerability of speaker verification to realistic voice spoofing,

    S. K. Erg ¨unay, E. Khoury, A. Lazaridis, and S. Marcel, “On the vulnerability of speaker verification to realistic voice spoofing,” in 2015 IEEE 7th international conference on biometrics theory, applications and systems (BTAS) . IEEE, 2015, pp. 1–6

    work page 2015

  8. [7]

    Generalization of spoofing countermeasures: A case study with asvspoof 2015 and btas 2016 corpora,

    D. Paul, M. Sahidullah, and G. Saha, “Generalization of spoofing countermeasures: A case study with asvspoof 2015 and btas 2016 corpora,” in 2017 IEEE international conference on acoustics, speech and signal processing (ICASSP) . IEEE, 2017, pp. 2047–2051

    work page 2015

Show all 34 references
  1. [8]

    An empirical study on channel effects for synthetic voice spoofing countermeasure systems,

    Y . Zhang, G. Zhu, F. Jiang, and Z. Duan, “An empirical study on channel effects for synthetic voice spoofing countermeasure systems,” arXiv preprint arXiv:2104.01320 , 2021

    2021

  2. [9]

    A study on data augmentation in voice anti-spoofing,

    A. Cohen, I. Rimon, E. Aflalo, and H. H. Permuter, “A study on data augmentation in voice anti-spoofing,” Speech Communication, vol. 141, pp. 56–67, 2022

    2022

  3. [10]

    Asvspoof 2019: Future horizons in spoofed and fake audio detection,

    M. Todisco, X. Wang, V . Vestman, M. Sahidullah, H. Delgado, A. Nautsch, J. Yamagishi, N. Evans, T. Kinnunen, and K. A. Lee, “Asvspoof 2019: Future horizons in spoofed and fake audio detection,” arXiv preprint arXiv:1904.05441 , 2019

    2019 arXiv

  4. [11]

    Asvspoof 2021: Towards spoofed and deepfake speech detection in the wild,

    X. Liu, X. Wang, M. Sahidullah, J. Patino, H. Delgado, T. Kinnunen, M. Todisco, J. Yamagishi, N. Evans, A. Nautsch et al., “Asvspoof 2021: Towards spoofed and deepfake speech detection in the wild,”IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 31, pp. ...

    2021

  5. [12]

    Audio deepfake detection: A survey,

    J. Yi, C. Wang, J. Tao, X. Zhang, C. Y . Zhang, and Y . Zhao, “Audio deepfake detection: A survey,” arXiv preprint arXiv:2308.14970 , 2023

    2023 arXiv

  6. [13]

    Audio anti-spoofing detection: A survey,

    M. Li, Y . Ahmadiadli, and X.-P. Zhang, “Audio anti-spoofing detection: A survey,” arXiv preprint arXiv:2404.13914 , 2024

    2024 arXiv

  7. [14]

    End-to-end spoofing detection with raw waveform cldnns,

    H. Dinkel, N. Chen, Y . Qian, and K. Yu, “End-to-end spoofing detection with raw waveform cldnns,” in 2017 IEEE international conference on acoustics, speech and signal processing (ICASSP) . IEEE, 2017, pp. 4860–4864

    2017

  8. [15]

    Recurrent convolutional structures for audio spoof and video deepfake detection,

    A. Chintha, B. Thai, S. J. Sohrawardi, K. Bhatt, A. Hickerson, M. Wright, and R. Ptucha, “Recurrent convolutional structures for audio spoof and video deepfake detection,” IEEE Journal of Selected Topics in Signal Processing , vol. 14, no. 5, pp. 1024–1037, 2020

    2020

  9. [16]

    Towards end-to-end synthetic speech detection,

    G. Hua, A. B. J. Teoh, and H. Zhang, “Towards end-to-end synthetic speech detection,” IEEE Signal Processing Letters , vol. 28, pp. 1265– 1269, 2021

    2021

  10. [17]

    End-to-end anti-spoofing with rawnet2,

    H. Tak, J. Patino, M. Todisco, A. Nautsch, N. Evans, and A. Larcher, “End-to-end anti-spoofing with rawnet2,” in ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2021, pp. 6369–6373

    2021

  11. [18]

    To- rawnet: improving rawnet with tcn and orthogonal regularization for fake audio detection,

    C. Wang, J. Yi, J. Tao, C. Zhang, S. Zhang, R. Fu, and X. Chen, “To- rawnet: improving rawnet with tcn and orthogonal regularization for fake audio detection,” arXiv preprint arXiv:2305.13701 , 2023

    2023 arXiv

  12. [19]

    End-to-end spectro-temporal graph attention networks for speaker ver- ification anti-spoofing and speech deepfake detection,

    H. Tak, J.-w. Jung, J. Patino, M. Kamble, M. Todisco, and N. Evans, “End-to-end spectro-temporal graph attention networks for speaker ver- ification anti-spoofing and speech deepfake detection,” arXiv preprint arXiv:2107.12710, 2021

    2021 arXiv

  13. [21]

    Raw differentiable architecture search for speech deepfake and spoofing detection,

    W. Ge, J. Patino, M. Todisco, and N. Evans, “Raw differentiable architecture search for speech deepfake and spoofing detection,” arXiv preprint arXiv:2107.12212, 2021

    2021 arXiv

  14. [22]

    Lever- aging positional-related local-global dependency for synthetic speech detection,

    X. Liu, M. Liu, L. Wang, K. A. Lee, H. Zhang, and J. Dang, “Lever- aging positional-related local-global dependency for synthetic speech detection,” in ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2023, pp. 1–5

    2023

  15. [23]

    Generalization of audio deepfake detection

    T. Chen, A. Kumar, P. Nagarsheth, G. Sivaraman, and E. Khoury, “Generalization of audio deepfake detection.” inOdyssey, 2020, pp. 132– 137

    2020

  16. [24]

    One-class learning towards synthetic voice spoofing detection,

    Y . Zhang, F. Jiang, and Z. Duan, “One-class learning towards synthetic voice spoofing detection,” IEEE Signal Processing Letters , vol. 28, pp. 937–941, 2021

    2021

  17. [25]

    Continual learning for fake audio detection,

    H. Ma, J. Yi, J. Tao, Y . Bai, Z. Tian, and C. Wang, “Continual learning for fake audio detection,” arXiv preprint arXiv:2104.07286 , 2021

    2021 arXiv

  18. [26]

    Rawboost: A raw data boosting and augmentation method applied to automatic speaker verification anti-spoofing,

    H. Tak, M. Kamble, J. Patino, M. Todisco, and N. Evans, “Rawboost: A raw data boosting and augmentation method applied to automatic speaker verification anti-spoofing,” in ICASSP 2022-2022 IEEE In- ternational Conference on Acoustics, Speech and Signal Processing (ICASSP). IEE...

    2022

  19. [27]

    Multi-dataset co-training with sharpness-aware optimization for audio anti-spoofing,

    H.-j. Shim, J.-w. Jung, and T. Kinnunen, “Multi-dataset co-training with sharpness-aware optimization for audio anti-spoofing,” arXiv preprint arXiv:2305.19953, 2023

    2023 arXiv

  20. [28]

    Deep neural networks are easily fooled: High confidence predictions for unrecognizable images,

    A. Nguyen, J. Yosinski, and J. Clune, “Deep neural networks are easily fooled: High confidence predictions for unrecognizable images,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 427–436

    2015

  21. [29]

    On calibration of modern neural networks,

    C. Guo, G. Pleiss, Y . Sun, and K. Q. Weinberger, “On calibration of modern neural networks,” in International conference on machine learning. PMLR, 2017, pp. 1321–1330

    2017

  22. [30]

    Evidential deep learning to quantify classification uncertainty,

    M. Sensoy, L. Kaplan, and M. Kandemir, “Evidential deep learning to quantify classification uncertainty,” Advances in neural information processing systems, vol. 31, 2018

    2018

  23. [31]

    Jsang, Subjective Logic: A formalism for reasoning under uncertainty

    A. Jsang, Subjective Logic: A formalism for reasoning under uncertainty. Springer Publishing Company, Incorporated, 2018

    2018

  24. [32]

    t-dcf: a detection cost function for the tandem assessment of spoofing countermeasures and automatic speaker verification,

    T. Kinnunen, K. A. Lee, H. Delgado, N. Evans, M. Todisco, M. Sahidul- lah, J. Yamagishi, and D. A. Reynolds, “t-dcf: a detection cost function for the tandem assessment of spoofing countermeasures and automatic speaker verification,” arXiv preprint arXiv:1804.09618 , 2018

    2018 arXiv

  25. [33]

    A comparative study on recent neural spoofing countermeasures for synthetic speech detection,

    X. Wang and J. Yamagishi, “A comparative study on recent neural spoofing countermeasures for synthetic speech detection,” arXiv preprint arXiv:2103.11326, 2021

    2021 arXiv

  26. [34]

    Optimiz- ing tandem speaker verification and anti-spoofing systems,

    A. Kanervisto, V . Hautam ¨aki, T. Kinnunen, and J. Yamagishi, “Optimiz- ing tandem speaker verification and anti-spoofing systems,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 30, pp. 477–488, 2021

    2021

Pith tools

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