# Generalized Fake Audio Detection via Deep Stable Learning

arXiv: https://arxiv.org/abs/2406.03237 · HTML: https://mlsanity.com/paper/2406.03237/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Jun 2024.

## Built on by (most cited first, 5 of 5)

- [Mixture of Experts Fusion for Fake Audio Detection Using Frozen wav2vec 2.0](https://mlsanity.com/paper/2409.11909/) (2024)
- [AASIST3: KAN-Enhanced AASIST Speech Deepfake Detection using SSL Features and Additional Regularization for the ASVspoof 2024 Challenge](https://mlsanity.com/paper/2408.17352/) (2024)
- [Does Current Deepfake Audio Detection Model Effectively Detect ALM-based Deepfake Audio?](https://mlsanity.com/paper/2408.10853/) (2024)
- [Temporal Variability and Multi-Viewed Self-Supervised Representations to Tackle the ASVspoof5 Deepfake Challenge](https://mlsanity.com/paper/2408.06922/) (2024)
- [Fake Speech Wild: Detecting Deepfake Speech on Social Media Platform](https://mlsanity.com/paper/2508.10559/) (2025)

## Abstract (arXiv)

Although current fake audio detection approaches have achieved remarkable success on specific datasets, they often fail when evaluated with datasets from different distributions. Previous studies typically address distribution shift by focusing on using extra data or applying extra loss restrictions during training. However, these methods either require a substantial amount of data or complicate the training process. In this work, we propose a stable learning-based training scheme that involves a Sample Weight Learning (SWL) module, addressing distribution shift by decorrelating all selected features via learning weights from training samples. The proposed portable plug-in-like SWL is easy to apply to multiple base models and generalizes them without using extra data during training. Experiments conducted on the ASVspoof datasets clearly demonstrate the effectiveness of SWL in generalizing different models across three evaluation datasets from different distributions.