Metric alignment with human understandability judgements on Known Corruptions BSL
1Pearson rBT2 Overall
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BT2 OverallRequirement=Reference-free (source text)2026.06 | 1 | 1 | |
| SVAERequirement=Reference-based (source sequence)2026.06 | 0.95 | 1 | |
| DTW-MPJPERequirement=Reference-based (source sequence)2026.06 | 0.91 | 0.5 | |
| SignCLIP - Multilingual sentenceMetric configuration=pose-to-text (p2t), Requirement=Reference-free (source text)2026.06 | 0.89 | 0.5 | |
| SignCLIP - Multilingual sentenceMetric configuration=pose-to-pose (p2p), Requirement=Reference-based (source sequence)2026.06 | 0.85 | 1 | |
| BT2 FidelityCategory=Per-component2026.06 | 0.8 | 1 | |
| SignCLIP - Multilingual segmentationMetric configuration=pose-to-text (p2t), Requirement=Reference-free (source text)2026.06 | 0.77 | 0.5 | |
| BT2 PhonologyCategory=Per-component2026.06 | 0.28 | 0.5 | |
| BT2 FluencyCategory=Per-component2026.06 | 0.28 | 0.5 | |
| SignCLIP - BSL sentenceMetric configuration=pose-to-text (p2t), Requirement=Reference-free (source text)2026.06 | 0.25 | 0.5 | |
| BLEURTRequirement=Reference-free (source text)2026.06 | 0.09 | 0.5 | |
| BLEU-4Requirement=Reference-free (source text)2026.06 | -0.19 | 0 | |
| ROUGERequirement=Reference-free (source text)2026.06 | -0.19 | 0 | |
| SignCLIP - BSL segmentationMetric configuration=pose-to-text (p2t), Requirement=Reference-free (source text)2026.06 | -0.25 | -0.5 | |
| SignCLIP - BSL segmentationMetric configuration=pose-to-pose (p2p), Requirement=Reference-based (source sequence)2026.06 | -0.55 | -0.5 | |
| SiBLEURequirement=Reference-based (source sequence)2026.06 | -0.58 | -0.5 | |
| SignCLIP - Multilingual segmentationMetric configuration=pose-to-pose (p2p), Requirement=Reference-based (source sequence)2026.06 | -0.59 | -0.5 | |
| ChrFRequirement=Reference-free (source text)2026.06 | -0.87 | -1 | |
| SignCLIP - BSL sentenceMetric configuration=pose-to-pose (p2p), Requirement=Reference-based (source sequence)2026.06 | -0.97 | -1 | |
| BT2 GrammarCategory=Per-component2026.06 | -1 | -1 |