Benchmarking Speech-to-Speech Translation Models 文章

ArXiv CS.CL2026-06-03NEWSen作者: Alkis Koudounas, Hayato Futami, Quentin Jodelet, Osamu Take, Shinji Watanabe, Emiru Tsunoo

摘要

arXiv:2606.03241v1 Announce Type: new Abstract: Speech-to-speech translation (S2ST) has advanced rapidly, but offline evaluation lacks a unified protocol: studies report non-overlapping metric subsets, preventing direct comparisons. We introduce COMPASS, a unified and reproducible benchmarking framework integrating 46 metrics across eight dimensions, and deploy it on 1,248 model-language configurations from FLEURS and CVSS, spanning cascaded and end-to-end architectures over ten language pairs. Architectures exhibit complementary strengths: best-vs-worst gaps exceed 30\% on naturalness and speaker preservation but remain within a few points on translation quality, so single-metric rankings systematically misrepresent system quality. Correlation filtering reduces 46 metrics to 10 per direction, with three axes requiring different metrics across X$\to$EN and EN$\to$X (e.g., TER/UTMOS vs. ChrF++/NISQA-MOS); these subsets preserve rankings (Spearman's $\rho>0.

相关事件查看全部 (1)

Benchmarking Speech-to-Speech Translation Models
2026-06-03PRODUCT_LAUNCH影响: MEDIUM

相关公司

暂无数据

相关人物

暂无数据