Conversational Retrieval on INSCIT
71.1Top-5 RecallWu et al. (2023)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Wu et al. (2023)2024.01 | 71.1 | — | |
| E5-unsupervised + Fine-tuneBackbone=E5, Query processing=History concatenation, Training protocol=Fine-tune, Training dataset=HumanAnnotatedConvQA2024.01 | 52.79 | 72.31 | |
| Dragon + RewriteBackbone=Dragon, Query processing=Rewrite, Training protocol=Unsupervised2024.01 | 52.79 | 68.92 | |
| Dragon + Fine-tune (SyntheticConvQA)Backbone=Dragon, Query processing=History concatenation, Training protocol=Fine-tune, Training dataset=SyntheticConvQA2024.01 | 47.41 | 67.13 | |
| Dragon + Fine-tuneBackbone=Dragon, Query processing=History concatenation, Training protocol=Fine-tune, Training dataset=HumanAnnotatedConvQA2024.01 | 43.43 | 64.94 | |
| DragonBackbone=Dragon, Training protocol=Unsupervised2024.01 | 27.49 | 46.22 | |
| E5-unsupervised + RewriteBackbone=E5, Query processing=Rewrite, Training protocol=Unsupervised2024.01 | 21.53 | 42.04 | |
| E5-unsupervisedBackbone=E5, Training protocol=Unsupervised2024.01 | 20.32 | 40.44 |