Segment-level Speech Translation Quality Estimation on IWSLT Metrics Shared Task 2026 (test)
26.7Kendall's Tau (en-de)HydraQE (xCOMET head)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HydraQE (xCOMET head)2026.06 | 26.7 | 32.9 | 29.8 | |
| HydraQE (DA head)2026.06 | 25.3 | 33.6 | 29.5 | |
| HydraQE (all heads avg.)2026.06 | 25.3 | 33.3 | 29.3 | |
| HydraQE (primary)2026.06 | 24.8 | 33.5 | 29.1 | |
| HydraQE (MetricX head)2026.06 | 24.3 | 32.9 | 28.6 | |
| CometKiwisource=organizers' results2026.06 | 23.7 | 33.3 | 28.5 | |
| SpeechQE2026.06 | 22.1 | 23.3 | 22.7 | |
| BLASER 2.0 QE2026.06 | 18.5 | 20.3 | 19.4 | |
| COMET (partial)2026.06 | 13.1 | 14.8 | 14 |