mlsanityData to 14 Oct 2026

Generalizable speech deepfake detection via meta-learned LoRA

arXiv 2502.10838 ↗First posted 15 Feb 2025eess.AS cs.LG cs.SDcited by 4

Not on the map. Why?

Abstract (arXiv)

Reliable detection of speech deepfakes (spoofs) must remain effective when the distribution of spoofing attacks shifts. We frame the task as domain generalization and show that inserting Low-Rank Adaptation (LoRA) adapters into every attention head of a self-supervised (SSL) backbone, then training only those adapters with Meta-Learning Domain Generalization (MLDG), yields strong zero-shot performance. The resulting model updates about 3.6 million parameters, roughly 1.1% of the 318 million updated in full fine-tuning, yet surpasses a fully fine-tuned counterpart on five of six evaluation corpora. A first-order MLDG loop encourages the adapters to focus on cues that persist across attack types, lowering the average EER from 8.84% for the fully fine-tuned model to 5.30% with our best MLDG-LoRA configuration. Our findings show that combining meta-learning with parameter-efficient adaptation offers an effective method for zero-shot, distribution-shift-aware speech deepfake detection.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 19 of its 25 references that are in the map. First posted on arXiv 15 Feb 2025.

Builds on

19 of 25 references.

References · 19

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  3. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,905
  4. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  5. 2018On First-Order Meta-Learning Algorithmscited by 1,018
  6. 2017Learning to Generalize: Meta-Learning for Domain Generalizationcited by 763
  7. 2019Domain Generalization via Model-Agnostic Learning of Semantic Featurescited by 407
  8. 2021FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Datasetcited by 181
  9. 2019ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detectioncited by 178
  10. 2021ASVspoof 2021: accelerating progress in spoofed and deepfake speech detectioncited by 161
  11. 2022Does Audio Deepfake Detection Generalize?cited by 135
  12. 2022Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentationcited by 117
7 more
  1. 2020Learning to Learn with Variational Information Bottleneck for Domain Generalizationcited by 102
  2. 2019STC Antispoofing Systems for the ASVspoof2019 Challengecited by 51
  3. 2021Continual Learning for Fake Audio Detectioncited by 27
  4. 2022On the Limitations of General Purpose Domain Generalisation Methodscited by 10
  5. 2023Low-rank Adaptation Method for Wav2vec2-based Fake Audio Detectioncited by 8
  6. 2023Adaptive Fake Audio Detection with Low-Rank Model Squeezingcited by 5
  7. 2024Meta-Learning Approaches for Improving Detection of Unseen Speech Deepfakescited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 2.

4 papers.

  1. 2025WaveSP-Net: Learnable Wavelet-Domain Sparse Prompt Tuning for Speech Deepfake Detectioncited by 5
  2. 2025Mixture of Low-Rank Adapter Experts in Generalizable Audio Deepfake Detectioncited by 2
  3. 2026CoReLoop: Parameter-Efficient Controlled Recurrent Refinement for Audio Deepfake Detectioncited by 2
  4. 2026Teffic-Audio: Tell Fact from Fictioncited by 1