Joint Optimization of ASV and CM tasks: BTUEF Team’s Submission for WildSpoof Challenge

Oğuzhan Kurnaz, Jagabandhu Mishra, Tomi Kinnunen, Cemal Hanilçi · 2026

Read Joint Optimization of ASV and CM tasks: BTUEF Team’s Submission for WildSpoof Challenge by Oğuzhan Kurnaz, Jagabandhu Mishra, Tomi Kinnunen, Cemal Hanilçi (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-2602-01722.txt.

Description

Oguzhan Kurnaz and colleagues jointly optimize speaker verification and spoofing countermeasures, reusing public systems through non linear fusion. Their modular framework explicitly models interaction between the two system types and trains with a condition dependent a-DCF loss. Experiments use ECAPA-TDNN and ReDimNet as verification extractors and SSL-AASIST as the countermeasure model, tested with and without fine tuning on WildSpoof data. Best performance comes from ReDimNet embeddings plus fine tuned SSL-AASIST representations, giving the lowest a-DCF on both progress and final sets.

Catalog record

CREATOROğuzhan Kurnaz, Jagabandhu Mishra, Tomi Kinnunen, Cemal Hanilçi
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE9793
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01722

Browse the full library · About the Tower of Babel