Lune

ACL Findings2026

When depth is redundant: Efficient transformer-based speech anti-spoofing

Hoan My Tran, Damien Lolive, Aghilas Sini, Arnaud Delhay, Pierre-Francois Marteau, David Guennec

Abstract

Detecting speech deepfakes is critical for protecting society against fraud, identity theft, and the misuse of modern speech synthesis technologies. Despite recent progress, existing countermeasures often exhibit limited generalization to unseen spoofing attacks, particularly in out-of-domain evaluation settings, even when achieving strong in-domain performance. Transformer architectures have become ubiquitous in anti-spoofing, serving both as feature extractors (e.g., wav2vec 2.0) and as classifiers. However, deep transformer stacks exhibit substantial representational redundancy across adjacent layers, with similarity increasing toward deeper layers. As a result, task-specific specialization is largely concentrated in the final layers, while shallow layers remain underutilized during fine-tuning. In this work, we analyze the layer-wise behavior of transformerbased classifiers for speech deepfake detection and propose a training strategy that explicitly aligns shallow and intermediate representations with those of the final transformer layer. By encouraging all layers to mimic the taskspecialized representation learned at depth, the model more effectively exploits early-layer features while preserving discriminative capacity in deeper layers. This design improves robustness to unseen spoofing attacks and enhances out-of-domain generalization. Extensive experiments across multiple benchmark datasets demonstrate consistent performance gains over strong baselines.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

Builds onCited byLeft to right by year; the more cited, the higher.

Builds on12

  1. wav2vec 2.0: A Framework for Self-Supervised Learning of Speech RepresentationsAlexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, Michael AuliNeurIPS 2020 · 9,451 citations
  2. Audio Deepfake Detection with Self-Supervised XLS-R and SLS ClassifierQishan Zhang, Shuangbing Wen, Tao HuACM MM 2024 · 54 citations
  3. Transferring Audio Deepfake Detection Capability across LanguagesZhongjie Ba, Qing Wen, Peng Cheng, Yuwei Wang et al.WWW 2023 · 34 citations
  4. SafeEar: Content Privacy-Preserving Audio Deepfake DetectionXinfeng Li, Kai Li, Yifan Zheng, Chen Yan et al.CCS 2024 · 26 citations
  5. Neural Collapse is Globally Optimal in Deep Regularized ResNets and TransformersPeter Súkeník, Christoph H. Lampert, Marco MondelliNeurIPS 2025 · 12 citations

Related papers

  1. A Unified Feature Mixture Framework for Joint Speech and Singing Deepfake DetectionAastha Sharma, Guangjing WangACL Findings 2026
  2. Alethia: a Foundational Encoder for Voice DeepfakesYi Zhu, Brahmi Dwivedi, Jayaram Raghuram, Surya KoppisettiICML 2026
  3. Learning Efficient Representations for Fake Speech DetectionNishant Subramani, Delip RaoAAAI 2020 · 46 citations
  4. SLIM: Style-Linguistics Mismatch Model for Generalized Audio Deepfake DetectionYi Zhu, Surya Koppisetti, Trang Tran, Gaurav BharajNeurIPS 2024 · 41 citations
  5. Face Forgery Detection via Symmetric TransformerLuchuan Song, Xiaodan Li, Zheng Fang, Zhenchao Jin et al.ACM MM 2022 · 20 citations
Dusk over the sea between two cliffs drawn in fine vertical lines