Semantic Textual Similarity on STS-B (test) using Kendall's Tau
76.5Kendall's Tau (STS-B)SGS
Evaluation Results
| Method | Links | |
|---|---|---|
| SGSHuman comparisons=600, Inferred comparisons=5352026.06 | 76.5 | |
| Random EloHuman comparisons=6002026.06 | 65.1 | |
| Active EloHuman comparisons=6002026.06 | 64.4 | |
| Active Elo (same budget)Human comparisons=6002026.06 | 63.5 | |
| RAIR-LaplacianHuman comparisons=6002026.06 | 62.5 | |
| PairS-MSHuman comparisons=6002026.06 | 17.9 |