HierCon: Hierarchical Contrastive Attention for Audio Deepfake Detection Conference: Proceedings of the ACM Web Conference 2026; April 13–17, 2026; Dubai, United Arab Emirates Proceedings of the ACM Web Conference 2026 (WWW ’26), April 13–17, 2026, Dubai, United Arab Emirates DOI: 10.1145/3774904.3792873 ISBN: 979-8-4007-2307-0/2026/04 CCS: Computing methodologies Speech recognition

Zhili Nicholas Liang, Soyeon Caren Han, Qizhou Wang, Christopher Leckie · 2026

Read HierCon: Hierarchical Contrastive Attention for Audio Deepfake Detection Conference: Proceedings of the ACM Web Conference 2026; April 13–17, 2026; Dubai, United Arab Emirates Proceedings of the ACM Web Conference 2026 (WWW ’26), April 13–17, 2026, Dubai, United Arab Emirates DOI: 10.1145/3774904.3792873 ISBN: 979-8-4007-2307-0/2026/04 CCS: Computing methodologies Speech recognition by Zhili Nicholas Liang, Soyeon Caren Han, Qizhou Wang, Christopher Leckie (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01032.txt.

Description

Zhili Nicholas Liang and co-authors build HierCon, a fake-voice spotter reading layers as a linked stack, not lone layers. Past spotters weigh layers alone, missing links across time and layer groups in made-up speech. HierCon weighs layers in rank order and trains with a margin rule pulling same-domain clips close and others apart. On two fake-voice sets it set new best error rates, beating lone-layer weighing by a wide margin. Its focus maps back the claim that linked layers travel better to new voices.

Catalog record

CREATORZhili Nicholas Liang, Soyeon Caren Han, Qizhou Wang, Christopher Leckie
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21475
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01032

Browse the full library · About the Tower of Babel