A conformer-based classifier for variable-length utterance processing in anti-spoofing
Open accessThe DOI matches its Crossref record; 40 of the 40 references it cites were found in a registry or an index.
Abstract
The success achieved by conformers in Automatic Speech Recognition (ASR) leads us to their application in other domains, such as spoofing detection for automatic speaker verification (ASV), where the conformer self-attention mechanism might effectively model and detect the artifacts introduced in spoofed speech signals. Also, conformers can naturally handle the variable duration of speech utterances. However, as with transformers, the conformer performance may degrade when trained with limited data. To address this issue, we propose utilizing conformers in conjunction with self-supervised learning, specifically leveraging a pre-trained model called wav2vec 2.0, which is pre-trained using a substantial amount of bonafide data. Our experimental results demonstrate that our proposed method achieves one of the best results in the recent ASVspoof 2021 logical access (LA) and deep fake (DF) databases.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Conformer [143] (EER increases from 3 . 28% to 94 . 82% on ASVSpoof21)”
From Deepfake detection across image, video, and audio: a comprehensive survey with empirical evaluation of generalization and robustness (2026) · cites this paper as [143]
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 1×
“The transformer architecture leverages self-attention to establish global dependencies between the input and output sequences [13].”
BackgroundSelf-citationExploiting Generative AI to Scale up Intelligent Tutoring Systems
Jakubův, J., Chvalovský, K., Goertzel, Z., Kaliszyk, C., Olšák, M., Piotrowski, B., Schulz, S., Suda, M., & Urban, J. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,658
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173 · cited 2×
“However, these models require large amounts of training data [16, 17, 18].”
BackgroundAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
Conformer: Convolution-augmented Transformer for Speech Recognition
Gulati, A., Qin, J., Chiu, C.-C., Parmar, N., Zhang, Y., Yu, J., Han, W., Wang, S., Zhang, Z., Wu, Y., et al. (2020) · Interspeech 2020 · doi:10.21437/interspeech.2020-3015 · Cited by 3,150
Conformer: Convolution-augmented Transformer for Speech Recognition
Gulati, A., Qin, J., Chiu, C.-C., Parmar, N., Zhang, Y., Yu, J., Han, W., Wang, S., Zhang, Z., Wu, Y., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.08100 · Cited by 3,150
HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units
Hsu, W.-N., Bolte, B., Tsai, Y.-H. H., Lakhotia, K., Salakhutdinov, R., & Mohamed, A. (2021) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2021.3122291 · Cited by 2,983
wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations
Baevski, A., Zhou, H., Mohamed, A., & Auli, M. (2020) · arXiv · doi:10.48550/arxiv.2006.11477 · Cited by 2,369
fairseq: A Fast, Extensible Toolkit for Sequence Modeling
Ott, M., Edunov, S., Baevski, A., Fan, A., Gross, S., Ng, N., Grangier, D., & Auli, M. (2019) · Proceedings of the 2019 Conference of the North · doi:10.18653/v1/n19-4009 · Cited by 2,344 · cited 1×
“The W2V model was implemented by using the Fairseq project toolkit [34].”
Methodwav2vec: Unsupervised Pre-Training for Speech Recognition
Schneider, S., Baevski, A., Collobert, R., & Auli, M. (2019) · Interspeech 2019 · doi:10.21437/interspeech.2019-1873 · Cited by 1,290
Cited by
XLSR-Mamba: A Dual-Column Bidirectional State Space Model for Spoofing Attack Detection
Xiao, Y., & Das, R. K. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3547861 · Cited by 103
“2 demonstrates the inference speed of XLSR-Mamba compared to recent XLSR-Conformer [19], measured using the real-time factor (RTF) on an NVIDIA A5000 GPU, with values averaged over 20 runs.”
InfluentialImproving Short Utterance Anti-Spoofing with Aasist2
Zhang, Y., Lu, J., Shang, Z., Wang, W., & Zhang, P. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10448049 · Cited by 64
Nes2Net: A Lightweight Nested Architecture for Foundation Model Driven Speech Anti-Spoofing
Liu, T., Truong, D.-T., Das, R. K., Lee, K. A., & Li, H. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3626963 · Cited by 69
“9.17 2024 Conformer [63] 8.42 2024 TCM [33] 7.79 2024 OCKD [70] 7.68 2024 SLS [35] 7.46 (8.87) 2024 Pascu et al. [74] - (7.2) 2025 Mamba [17] 6.71 2025 WaveSpec [72] 6.58 2025 LSR+LSA [71] 5.92 2025 LSR+LSA [71] ※ 5.54 - Proposed Nes2Net 5.80 (7.06) - Proposed Nes2Net-X 5.52 (6.60) ※ : with extra…”
InfluentialAdvancements in AI-Generated Content Forensics: A Systematic Literature Review
Xu, Q., Mu, W., Li, J., Sun, T., & Jiang, X. (2025) · ACM Computing Surveys · doi:10.1145/3760526 · Cited by 26
“More recently, advanced architectures like the Conformer model [185], known for its ability to capture multi-scale features, have been adopted and enhanced with self-attention mechanisms to further boost performance.”
Martín-Doñas, J. M., Álvarez, A., Rosello, E., Gomez, A. M., & Peinado, A. M. (2024) · Interspeech 2024 · doi:10.21437/interspeech.2024-942 · Cited by 18
“…the development of deepfake detectors built upon these SSL architectures, focusing on the following aspects: improving the downstream network [13, 14], using different upstreams such as WavLM [15] or Whisper [16], exploring multi-corpus with domain-invariant training [17], or…”
MethodInfluentialSelf-citationTrident of Poseidon: A Generalized Approach for Detecting Deepfake Voices
Doan, T.-P., Dinh-Xuan, H., Ryu, T., Kim, I., Lee, W., Hong, K., & Jung, S. (2024) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3658644.3690311 · Cited by 26
Xuan, X., Zhu, Z., Zhang, W., Lin, Y.-C., & Kinnunen, T. (2025) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru65441.2025.11434679 · Cited by 22
“Figure 4 compares the inference speed of Fake-Mamba(L) with the latest XLSR-Conformer [59] model in terms of average Real-Time Factor (RTF) over 100 runs on the same hardware.”
InfluentialKim, I., Doan, T.-P., Hong, S., & Jung, S. (2025) · IEEE Access · doi:10.1109/access.2025.3635633 · Cited by 11
Multi-Scale Information Aggregation for Spoofing Detection
Li, C., Wan, Y., Yang, F., & Yang, J. J. (2022) · EURASIP Journal on Audio Speech and Music Processing · doi:10.1186/s13636-024-00379-x · Cited by 8
Hao, Y., Xu, M., Chen, Y., Liu, Y., He, L., Fang, L., & Liu, L. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889337 · Cited by 11
Room Impulse Responses Help Attackers to Evade Deep Fake Detection
Luong, H.-T., Truong, D.-T., Lee, K. A., & Chng, E. S. (2024) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt61566.2024.10832177 · Cited by 8
One-class network leveraging spectro-temporal features for generalized synthetic speech detection
Ye, J., Yan, D., Fu, S., Ma, B., & Xia, Z. (2025) · Speech Communication · doi:10.1016/j.specom.2025.103200 · Cited by 9
Robust Audio Deepfake Detection using Ensemble Confidence Calibration
Kwok, C. Y., Truong, D.-T., & Yip, J. Q. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889972 · Cited by 9
“We utilize the XLSR-Conformer [10] architecture as our baseline.”
InfluentialMulti-level SSL Feature Gating for Audio Deepfake Detection
Tran, H. M., Lolive, D., Sini, A., Delhay, A., Marteau, P.-F., & Guennec, D. (2025) · Proceedings of the 33rd ACM International Conference on Multimedia · doi:10.1145/3746027.3754568 · Cited by 4
“This has shown effectiveness in Automatic Speech Recognition (ASR) as well as in SSD [40, 49].”
Hao, Y., Chen, Y., Xu, M., Zhan, J., He, L., Fang, L., Fang, S., & Liu, L. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11228261 · Cited by 8
AFSS: Artifact-Focused Self-Synthesis for Mitigating Bias in Audio Deepfake Detection
Nguyen-Le, H.-S., Nguyen-Thanh, H.-C., Le‐Khac, N., Nguyen, D.-T., & Nguyen-Le, H.-H. (2026) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn69854.2026.11728808 · Cited by 3
“We compare our method with 5 baselines, including AASIST-B [10], AASIST-L [10], SCL [17], Conformer-based detector [18], and SSL [20].”
Comprehensive Validation of Bridge Module and EBM Loss for One-Class Audio Deepfake Detection
Kim, S.-B., Shin, H.-S., Heo, J., Lim, C.-Y., Koo, K.-W., Son, J., Jung, S., & Yu, H.-J. (2026) · IEEE Access · doi:10.1109/access.2026.3680160 · Cited by 3
Towards Real-Time Audio Deepfake Detection in Resource-Limited Environments
Dinh-Xuan, H., Doan, T.-P., Han, S., Hong, K., & Jung, S. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-7005-5_28 · Cited by 4
XLSR-Kanformer: A KAN-Intergrated model for Synthetic Speech Detection
Dat, P. T., & Dat, T. H. (2025) · IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) · doi:10.1109/avss65446.2025.11149956 · Cited by 9
“In particular, [10] utilizes an SSL model as a feature extractor to derive meaningful representations from raw waveform inputs.”
Viakhirev, I., Borodin, K., Gorodnichev, M. G., & Mkrtchian, G. (2026) · Mathematics · doi:10.3390/math14020381 · Cited by 2
“…framework is, by construction, model-agnostic and can in principle be applied to other mainstream multi-branch anti-spoofing systems, including RawNet-style encoders [27], graph-based countermeasures [24], and recent convolutional–transformer hybrids [28] used in ASVspoof challenges [29].”
RIRplay: Generation of a Replay Stereo Corpus for Voice Biometrics Anti-Spoofing
Sánchez-Valera, J. C., Peinado, A. M., Martín-Doñas, J. M., Gomez-Alanis, A., Gomez, A. M., & Todisco, M. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3684815 · Cited by 4
Self-citationFine-Grained Frame Modeling in Multi-Head Self-Attention for Speech Deepfake Detection
Dat, P. T., Truong, D.-T., Hoang, L.-V., & Trang, N. T. T. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464918 · Cited by 4
Recursive Feature Learning from Pre-Trained Models for Spoofing Speech Detection
Guan, Y., Ai, Y., Li, Z., Peng, S., & Guo, W. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888985 · Cited by 3
Nguyen-Le, H.-H., Tran, V.-T., Nguyen, D.-T., & Le-Khac, N.-A. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11608-4 · Cited by 28
Faking Fluent: Unveiling the Achilles’ Heel of Multilingual Deepfake Detection
Ranjan, R., Dutta, B., Mayank Vatsa, A., & Singh, R. (2024) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744454 · Cited by 6
KAN You Hear the Truth? Audio Deepfake Detection with Kolmogorov–Arnold Networks
Tran, H. M., Lolive, D., Guennec, D., Sini, A., Delhay, A., & Marteau, P.-F. (2025) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco63237.2025.11226534 · Cited by 4
“As with the original design [10], a learnable classification token X CLS ∈ R C is prepended to form the X seq = [ X CLS , X SSL ] ∈ R ( T +1) × C before being fed into the Conformer model consisting of L blocks.”
Audio Deepfake Detection at the First Greeting: "Hi!"
Shi, H., Shi, X., Dogan, S., Huang, T., & Zhang, Y. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463587 · Cited by 1
“Artificially extending inputs at inference time is also impractical for latency-sensitive applications [12,13].”
Robust Deepfake Audio Detection via Multi-Level Intermediate Feature Fusion
Zhao, J. S., Zhao, J. S., Zhou, Y., Zheng, P., & Du, Y. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462275 · Cited by 1
Assessing the Impact of Speaker Identity in Speech Spoofing Detection
Dao, A.-T., Matrouf, D., & Evans, N. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462102 · Cited by 5
Dao, A.-T., Matrouf, D., Rouvier, M., & Evans, N. (2026) · International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf68042.2026.11558167 · Cited by 2
“Rosell´o et al. [7] utilized Con-former architectures to capture both local and global speech patterns, achieving state-of-the-art performance using the same datasets.”
PULMO: Precise utterance-level modeling for speech anti-spoofing
Yoon, S.-H. (2024) · Applied Acoustics · doi:10.1016/j.apacoust.2024.110221 · Cited by 2
Enhancing Deepfake Detection: A Study Using WavLM and Advanced RawBoost Augmentation Techniques
Tri, N., Hoang, L. N., Viet, P. T., & Trung, K. P. (2024) · RIVF International Conference on Computing and Communication Technologies (RIVF) · doi:10.1109/rivf64335.2024.11009068
ConMSDMamba: Multi-Scale Dilated Mamba Based on Conformer for Speech Emotion Recognition
Qian, G., Lei, Z., Liu, S., Liu, C., & Jiang, A. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3632762
Zhong, H., Wang, W., & Zhang, J. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11227892
Yu, G., Guo, W., Zhang, J., & Zhang, Z. (2025) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc65261.2025.11249219 · Cited by 2
“Recently, pre-trained models (PTMs), e.g., Wav2Vec2 [6], HuBERT [7], WavLM[8], have significantly boosted the performance of many downstream speech processing tasks, including speech recognition, speaker recognition and SSD [9], [10].”
InfluentialAudio Deepfake Detection via Dual Branch Classifier with Self-Supervised Pre-Trained Model
Cao, Y., Gao, L., & Mao, Q. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-5382-2_25
Audio-Visual Cross-Attention for Improved Deepfake Video Detection and Forgery Localization
Jalleli, O., Zhu, Y. M., & Falk, T. H. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11342953
Enhanced Spoofing Speech Detection with Self-supervised Feature Extraction and Multi-expert Models
Long, Y., Wei, G., Zhang, J., Guo, B., Wang, W., & Liu, G. (2026) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-032-21168-2_7
Kim, I., Seo, J., Park, S., Doan, T.-P., & Jung, S. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464242
Orthogonal Disentanglement for Generalizable Audio Deepfake Detection
Park, J. H., Lee, S., & Kwon, J. (2026) · International Conference on Information Networking · doi:10.1109/icoin68469.2026.11480561
“Let h ∈ R d denote the latent embedding obtained from the encoder [14].”
Comparative Eval. of SSL Feature Extractors & Classifiers for DFAD
Park, S.-B., & Song, H.-M. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-6857-4_24
Interpreting SSL Representations for Spoof Detection: A WavLM Study
Mallat, M., Panariello, M., Todisco, M., Evans, N., & Larcher, A. (2026) · International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf68042.2026.11558152
“Neural spoofing countermeasures (CMs) now span a broad range of back-ends, including graph-attention architectures such as AASIST, conformer-based anti-spoofing classifiers, and fused/ensemble systems [6], [7], [1], [8], [9].”
MSLA-XLS-R: Enhancing hierarchical SSL representations for audio deepfake detection
Meng, H., Tang, Y., Sun, Y., Deng, Y., Li, Z., & Li, Y. (2026) · Computer Speech & Language · doi:10.1016/j.csl.2026.102019
Domain-Invariant Representation Learning for Generalizable Deepfake Speech Detection
Zhang, X., Zhu, G., Fu, X., Jiang, W., Xiang, T., Li, Q., & Wu, L. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3710772
Robocalls: A Worldwide or US-only Problem? Analyzing Spam and Fraud in International Phone CallsPreprint
Altwlkany, K., Merćep, A., Đuričić, T., Kapetanović, A. L., & Lacić, E. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7079425
FePS-Net: A Frequency Enhanced Prosody–Style Fusion Network for Deepfake Speech Detection
Zhao, Y., Wei, L., Jin, X., & Wang, G. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3438-7_30
Exploiting Neural Speech Enhancement Models to Undermine Voice Watermarking
Gomez, A. M., Martín-Doñas, J. M., & Peinado, A. M. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3718252
Self-citationImproving Audio Anti-Spoofing with FLA and Local-global Feature Fusion
Sang, J., & Huang, Z. (2026) · International Conference on Electronic Engineering and Informatics (EEI) · doi:10.1109/eei70303.2026.11640547 · Cited by 1
Zhang, G., Liu, S., Zhang, D., & Tang, L. (2026) · Applied Sciences · doi:10.3390/app16178413 · Cited by 1
Audio Deepfake Detection: A Comprehensive Survey of Generation, Datasets, and Detection Strategies
Abbas, S. S., Al-Obaidi, S., & Al-Sultan, A. Y. (2026) · International Journal of Safety and Security Engineering · doi:10.18280/ijsse.160718
Nguyen, T. A., Dinh-Xuan, H., Doan, T.-P., & Jung, S. (2026) · IEEE Access · doi:10.1109/access.2026.3732510
Roselló, E., López‐Espejo, I., Peinado, A. M., & Gomez, A. M. (2026) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2026.3739196
Self-citationQuantum Kernel-Based Support Vector Machine for Deepfake Audio Detection
Thakur, S., Singh, M., & Modi, S. (2026) · International Conference on Innovative Sustainable Computational Technologies (CISCT) · doi:10.1109/cisct69589.2026.11726908
“Other studies have proposed conformer-based detection models [10], self-supervised domain-invariant representations [11], Whisper-based feature extraction [12], mono-to-stereo conversion-based models [13], and complex-valued neural networks [14].”
GLC-Conformer: Global-to-Local Calibration for Audio Anti-Spoofing
Yang, M., Zheng, K., Yao, Y., Gao, J., & Zhao, Y. (2026) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn69854.2026.11729791
Metadata
- Authors
- Rosello, E., Gomez-Alanis, A., Gomez, A. M., & Peinado, A.
- Affiliations
- Universidad de Granada (ES)
- Venue
- INTERSPEECH 2023
- Publisher
- ISCA
- Published
- 2023-08-20
- Pages
- 5281-5285
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W4385822395
- Cited by
- 47
- References
- 35
- Citations by year
- 2026: 24 · 2025: 16 · 2024: 6
- FWCI
- 8.14
- Type
- conference paper
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- other-oa
- Locations
- Interspeech (USB) (publishedVersion)
- Institutional Repository of the University of Granada (University of Granada) (submittedVersion, other-oa)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Biometric Identification and Security
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Keywords
- conformational isomers, anti-spoofing, speaker verification, self-supervised learning, wav2vec 2.0, self-attention, deepfake detection, ASVspoof 2021
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.