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

arXiv: https://arxiv.org/abs/2512.21702 · HTML: https://mlsanity.com/paper/2512.21702/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 38 references that are in the map.
- First posted on arXiv 25 Dec 2025.

## Builds on (most cited first, 27 of 27)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://mlsanity.com/paper/2110.13900/) (2021)
- [Common Voice: A Massively-Multilingual Speech Corpus](https://mlsanity.com/paper/1912.06670/) (2019)
- [AST: Audio Spectrogram Transformer](https://mlsanity.com/paper/2104.01778/) (2021)
- [PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition](https://mlsanity.com/paper/1912.10211/) (2019)
- [Unsupervised Cross-lingual Representation Learning for Speech Recognition](https://mlsanity.com/paper/2006.13979/) (2020)
- [ASVspoof 2019: Future Horizons in Spoofed and Fake Audio Detection](https://mlsanity.com/paper/1904.05441/) (2019)
- [ASVspoof 2021: accelerating progress in spoofed and deepfake speech detection](https://mlsanity.com/paper/2109.00537/) (2021)
- [Does Audio Deepfake Detection Generalize?](https://mlsanity.com/paper/2203.16263/) (2022)
- [Audio Deepfake Detection: A Survey](https://mlsanity.com/paper/2308.14970/) (2023)
- [Audio Deepfake Detection with Self-Supervised WavLM and Multi-Fusion Attentive Classifier](https://mlsanity.com/paper/2312.08089/) (2023)
- [RawBMamba: End-to-End Bidirectional State Space Model for Audio Deepfake Detection](https://mlsanity.com/paper/2406.06086/) (2024)
- [ASSERT: Anti-Spoofing with Squeeze-Excitation and Residual neTworks](https://mlsanity.com/paper/1904.01120/) (2019)
- [A robust audio deepfake detection system via multi-view feature](https://mlsanity.com/paper/2403.01960/) (2024)
- [Human Perception of Audio Deepfakes](https://mlsanity.com/paper/2107.09667/) (2021)
- [What to Remember: Self-Adaptive Continual Learning for Audio Deepfake Detection](https://mlsanity.com/paper/2312.09651/) (2023)
- [Defense Against Adversarial Attacks on Audio DeepFake Detection](https://mlsanity.com/paper/2212.14597/) (2022)
- [WavLM model ensemble for audio deepfake detection](https://mlsanity.com/paper/2408.07414/) (2024)
- [SpecRNet: Towards Faster and More Accessible Audio DeepFake Detection](https://mlsanity.com/paper/2210.06105/) (2022)
- [Trusted Media Challenge Dataset and User Study](https://mlsanity.com/paper/2201.04788/) (2022)
- [Region-Based Optimization in Continual Learning for Audio Deepfake Detection](https://mlsanity.com/paper/2412.11551/) (2024)
- [Audio Spoofing Verification using Deep Convolutional Neural Networks by Transfer Learning](https://mlsanity.com/paper/2008.03464/) (2020)
- [Continuous Learning of Transformer-based Audio Deepfake Detection](https://mlsanity.com/paper/2409.05924/) (2024)
- [SynHate: Detecting Hate Speech in Synthetic Deepfake Audio](https://mlsanity.com/paper/2506.06772/) (2025)
- [BanglaFake: Constructing and Evaluating a Specialized Bengali Deepfake Audio Dataset](https://mlsanity.com/paper/2505.10885/) (2025)
- [Multimodal Zero-Shot Framework for Deepfake Hate Speech Detection in Low-Resource Languages](https://mlsanity.com/paper/2506.08372/) (2025)

## 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.