mlsanityData to 7 Oct 2026

Zero-Shot to Zero-Lies: Detecting Bengali Deepfake Audio through Transfer Learning

arXiv 2512.21702 ↗First posted 25 Dec 2025cs.SD cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

The rapid growth of speech synthesis and voice conversion systems has made deepfake audio a major security concern. Bengali deepfake detection remains largely unexplored. In this work, we study automatic detection of Bengali audio deepfakes using the BanglaFake dataset. We evaluate zeroshot inference with several pretrained models. These include Wav2Vec2-XLSR-53, Whisper, PANNsCNN14, WavLM and Audio Spectrogram Transformer. Zero-shot results show limited detection ability. The best model, Wav2Vec2-XLSR-53, achieves 53.80% accuracy, 56.60% AUC and 46.20% EER. We then f ine-tune multiple architectures for Bengali deepfake detection. These include Wav2Vec2-Base, LCNN, LCNN-Attention, ResNet18, ViT-B16 and CNN-BiLSTM. Fine-tuned models show strong performance gains. ResNet18 achieves the highest accuracy of 79.17%, F1 score of 79.12%, AUC of 84.37% and EER of 24.35%. Experimental results confirm that fine-tuning significantly improves performance over zero-shot inference. This study provides the first systematic benchmark of Bengali deepfake audio detection. It highlights the effectiveness of f ine-tuned deep learning models for this low-resource language.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 27 of its 38 references that are in the map. First posted on arXiv 25 Dec 2025.

Builds on

27 of 38 references.

References · 27

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 3,972
  2. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  3. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 1,609
  4. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,358
  5. 2021AST: Audio Spectrogram Transformercited by 634
  6. 2019PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognitioncited by 563
  7. 2020Unsupervised Cross-lingual Representation Learning for Speech Recognitioncited by 496
  8. 2019ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detectioncited by 173
  9. 2021ASVspoof 2021: accelerating progress in spoofed and deepfake speech detectioncited by 157
  10. 2022Does Audio Deepfake Detection Generalize?cited by 129
  11. 2023Audio Deepfake Detection: A Surveycited by 49
  12. 2023Audio Deepfake Detection with Self-Supervised WavLM and Multi-Fusion Attentive Classifiercited by 26
15 more
  1. 2024RawBMamba: End-to-End Bidirectional State Space Model for Audio Deepfake Detectioncited by 26
  2. 2019ASSERT: Anti-Spoofing with Squeeze-Excitation and Residual neTworkscited by 23
  3. 2024A robust audio deepfake detection system via multi-view featurecited by 19
  4. 2021Human Perception of Audio Deepfakescited by 15
  5. 2023What to Remember: Self-Adaptive Continual Learning for Audio Deepfake Detectioncited by 15
  6. 2022Defense Against Adversarial Attacks on Audio DeepFake Detectioncited by 12
  7. 2024WavLM model ensemble for audio deepfake detectioncited by 12
  8. 2022SpecRNet: Towards Faster and More Accessible Audio DeepFake Detectioncited by 9
  9. 2022Trusted Media Challenge Dataset and User Studycited by 7
  10. 2024Region-Based Optimization in Continual Learning for Audio Deepfake Detectioncited by 5
  11. 2020Audio Spoofing Verification using Deep Convolutional Neural Networks by Transfer Learningcited by 3
  12. 2024Continuous Learning of Transformer-based Audio Deepfake Detectioncited by 3
  13. 2025SynHate: Detecting Hate Speech in Synthetic Deepfake Audiocited by 3
  14. 2025BanglaFake: Constructing and Evaluating a Specialized Bengali Deepfake Audio Datasetcited by 1
  15. 2025Multimodal Zero-Shot Framework for Deepfake Hate Speech Detection in Low-Resource Languagescited by 1

Built on by

Not cited yet.