Answer Sentence Selection on Alexa Virtual Assistant traffic Sample 1 (test)
71.26Prec@1TANDA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TANDABackbone=RoBERTa, Model Size=Large, Transfer Sequence=ASNQ → NAD2019.11 | 71.26 | 0.68 | 0.805 | |
| RoBERTaModel Size=Large, Training Strategy=Fine-tuning on NAD2019.11 | 70.81 | 0.654 | 0.796 | |
| TANDABackbone=RoBERTa, Model Size=Base, Transfer Sequence=ASNQ → NAD2019.11 | 65.59 | 0.623 | 0.757 | |
| RoBERTaModel Size=Large, Training Strategy=Zero-shot from ASNQ2019.11 | 64.37 | 0.627 | 0.75 | |
| TANDABackbone=BERT, Model Size=Large, Transfer Sequence=ASNQ → NAD2019.11 | 61.54 | 0.607 | 0.725 | |
| RoBERTaModel Size=Base, Training Strategy=Fine-tuning on NAD2019.11 | 59.11 | 0.563 | 0.699 | |
| TANDABackbone=BERT, Model Size=Base, Transfer Sequence=ASNQ → NAD2019.11 | 58.7 | 0.585 | 0.703 | |
| RoBERTaModel Size=Base, Training Strategy=Zero-shot from ASNQ2019.11 | 58.7 | 0.587 | 0.707 | |
| BERTModel Size=Large, Training Strategy=Zero-shot from ASNQ2019.11 | 57.49 | 0.552 | 0.686 | |
| BERTModel Size=Base, Training Strategy=Zero-shot from ASNQ2019.11 | 55.06 | 0.557 | 0.677 | |
| BERTModel Size=Large, Training Strategy=Fine-tuning on NAD2019.11 | 53.85 | 0.537 | 0.671 | |
| BERTModel Size=Base, Training Strategy=Fine-tuning on NAD2019.11 | 49.8 | 0.506 | 0.638 |