mlsanityData to 14 Oct 2026

Improving Generalization for AI-Synthesized Voice Detection

arXiv 2412.19279 ↗First posted 26 Dec 2024cs.SD cs.LG eess.AScited by 6

Not on the map. Why?

Abstract (arXiv)

AI-synthesized voice technology has the potential to create realistic human voices for beneficial applications, but it can also be misused for malicious purposes. While existing AI-synthesized voice detection models excel in intra-domain evaluation, they face challenges in generalizing across different domains, potentially becoming obsolete as new voice generators emerge. Current solutions use diverse data and advanced machine learning techniques (e.g., domain-invariant representation, self-supervised learning), but are limited by predefined vocoders and sensitivity to factors like background noise and speaker identity. In this work, we introduce an innovative disentanglement framework aimed at extracting domain-agnostic artifact features related to vocoders. Utilizing these features, we enhance model learning in a flat loss landscape, enabling escape from suboptimal solutions and improving generalization. Extensive experiments on benchmarks show our approach outperforms state-of-the-art methods, achieving up to 5.12% improvement in the equal error rate metric in intra-domain and 7.59% in cross-domain evaluations.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 53 of its 75 references that are in the map. First posted on arXiv 26 Dec 2024.

Builds on

53 of 75 references.

References · 53

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  3. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  4. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  5. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  6. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 2,010
  7. 2016WaveNet: A Generative Model for Raw Audiocited by 1,722
  8. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 1,657
  9. 2020Sharpness-Aware Minimization for Efficiently Improving Generalizationcited by 1,417
  10. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesiscited by 1,354
  11. 2018Learning deep representations by mutual information estimation and maximizationcited by 1,273
  12. 2020HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesiscited by 1,154
38 more
  1. 2020FastSpeech 2: Fast and High-Quality End-to-End Text to Speechcited by 717
  2. 2018MINE: Mutual Information Neural Estimationcited by 683
  3. 2020WaveGrad: Estimating Gradients for Waveform Generationcited by 622
  4. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 601
  5. 2021Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speechcited by 561
  6. 2019MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesiscited by 354
  7. 2019Parallel WaveGAN: A fast waveform generation model based on generative adversarial networks with multi-resolution spectrogramcited by 340
  8. 2018WaveGlow: A Flow-based Generative Network for Speech Synthesiscited by 309
  9. 2018Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesiscited by 282
  10. 2018Efficient Neural Audio Synthesiscited by 219
  11. 2021FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Datasetcited by 181
  12. 2021AASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networkscited by 180
  13. 2023UCF: Uncovering Common Features for Generalizable Deepfake Detectioncited by 180
  14. 2020End-to-end anti-spoofing with RawNet2cited by 164
  15. 2021ASVspoof 2021: accelerating progress in spoofed and deepfake speech detectioncited by 161
  16. 2021WaveFake: A Data Set to Facilitate Audio Deepfake Detectioncited by 78
  17. 2020End-to-End Adversarial Text-to-Speechcited by 77
  18. 2018ClariNet: Parallel Wave Generation in End-to-End Text-to-Speechcited by 70
  19. 2024Preserving Fairness Generalization in Deepfake Detectioncited by 70
  20. 2018Learning latent representations for style control and transfer in end-to-end speech synthesiscited by 61
  21. 2021StarGANv2-VC: A Diverse, Unsupervised, Non-parallel Framework for Natural-Sounding Voice Conversioncited by 58
  22. 2018FloWaveNet : A Generative Flow for Raw Audiocited by 54
  23. 2019STC Antispoofing Systems for the ASVspoof2019 Challengecited by 51
  24. 2013Equitability, mutual information, and the maximal information coefficientcited by 45
  25. 2023AI-Synthesized Voice Detection Using Neural Vocoder Artifactscited by 41
  26. 2020Wave-Tacotron: Spectrogram-free end-to-end text-to-speech synthesiscited by 38
  27. 2017Wasserstein CNN: Learning Invariant Features for NIR-VIS Face Recognitioncited by 37
  28. 2022SAMO: Speaker Attractor Multi-Center One-Class Learning for Voice Anti-Spoofingcited by 18
  29. 2023Can large-scale vocoded spoofed data improve speech spoofing countermeasure with a self-supervised front end?cited by 17
  30. 2020Speaker Conditional WaveRNN: Towards Universal Neural Vocoder for Unseen Speaker and Recording Conditionscited by 14
  31. 2020EfficientTTS: An Efficient and High-Quality Text-to-Speech Architecturecited by 14
  32. 2020An explainability study of the constant Q cepstral coefficient spoofing countermeasure for automatic speaker verificationcited by 13
  33. 2023Do You Remember? Overcoming Catastrophic Forgetting for Fake Audio Detectioncited by 11
  34. 2023Detection of Cross-Dataset Fake Audio Based on Prosodic and Pronunciation Featurescited by 10
  35. 2023Single and Multi-Speaker Cloned Voice Detection: From Perceptual to Learned Featurescited by 10
  36. 2019Ensemble Models for Spoofing Detection in Automatic Speaker Verificationcited by 8
  37. 2021Many-to-Many Voice Conversion based Feature Disentanglement using Variational Autoencodercited by 7
  38. 2023DSVAE: Interpretable Disentangled Representation for Synthetic Speech Detectioncited by 6

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year

Citing papers per year: 2025: 6.

6 papers.

  1. 2025Improving Generalization of Medical Image Registration Foundation Modelcited by 6
  2. 2025Robust AI-Generated Face Detection with Imbalanced Datacited by 5
  3. 2025Rethinking Individual Fairness in Deepfake Detectioncited by 4
  4. 2025AUDETER: A Large-scale Dataset for Deepfake Audio Detection in Open Worldscited by 0
  5. 2025Can Current Detectors Catch Face-to-Voice Deepfake Attacks?cited by 0
  6. 2025SynthGuard: An Open Platform for Detecting AI-Generated Multimedia with Multimodal LLMscited by 0