Response Selection on ConvAI2 (test)
87.9R@20Cross-encoder
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Cross-encoderPre-training Data=Reddit, Pre-training Source=Our pre-training2019.04 | 87.9 | — | |
| Poly-encoder 360Pre-training Data=Reddit, Pre-training Source=Our pre-training, m context vectors=3602019.04 | 86.8 | — | |
| Poly-encoder 64Pre-training Data=Reddit, Pre-training Source=Our pre-training, m context vectors=642019.04 | 86.5 | — | |
| Poly-encoder 16Pre-training Data=Reddit, Pre-training Source=Our pre-training, m context vectors=162019.04 | 86.3 | — | |
| Cross-encoderPre-training Data=Toronto Books + Wikipedia, Pre-training Source=Our pre-training2019.04 | 84.9 | — | |
| Cross-encoderPre-training Data=Toronto Books + Wikipedia, Pre-training Source=Devlin et al., 20192019.04 | 84.8 | — | |
| Bi-encoderPre-training Data=Reddit, Pre-training Source=Our pre-training2019.04 | 84.8 | — | |
| Poly-encoder 360Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Our pre-training, m context vectors=3602019.04 | 83.8 | — | |
| Poly-encoder 64Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Devlin et al., 2019, m context vectors=642019.04 | 83.7 | — | |
| Poly-encoder 360Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Devlin et al., 2019, m context vectors=3602019.04 | 83.7 | — | |
| Poly-encoder 64Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Our pre-training, m context vectors=642019.04 | 83.3 | — | |
| Poly-encoder 16Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Devlin et al., 2019, m context vectors=162019.04 | 83.2 | — | |
| Poly-encoder 16Pre-training Data=Toronto Books + Wikipedia, Pre-training Source=Our pre-training, m context vectors=162019.04 | 82.7 | — | |
| Bi-encoderPre-training Data=Toronto Books + Wikipedia, Pre-training Source=Our pre-training2019.04 | 82 | — | |
| Bi-encoderPre-training Data=Toronto Books + Wikipedia, Pre-training Source=Devlin et al., 20192019.04 | 81.7 | — | |
| Wolf et al.2019.04 | 80.7 | — | |
| Bi-encoderInitialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 81.7 | |
| Bi-encoderInitialization=This paper, Pre-training corpus=Reddit2019.04 | — | 84.8 | |
| Cross-encoderInitialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 84.8 | |
| Cross-encoderInitialization=This paper, Pre-training corpus=Reddit2019.04 | — | 87.9 | |
| Hugging Face2019.04 | — | 80.7 | |
| Poly-encoderm=16, Mechanism=First-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 83.9 | |
| Poly-encoderm=16, Mechanism=Learnt-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 83.2 | |
| Poly-encoderm=64, Mechanism=First-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 84.2 | |
| Poly-encoderm=64, Mechanism=Learnt-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 83.7 | |
| Poly-encoderm=360, Mechanism=First-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 84.6 | |
| Poly-encoderm=360, Mechanism=Learnt-m, Initialization=Devlin et al., 2019, Pre-training corpus=Toronto Books + Wikipedia2019.04 | — | 83.7 | |
| Poly-encoderm=16, Mechanism=First-m, Initialization=This paper, Pre-training corpus=Reddit2019.04 | — | 86.4 | |
| Poly-encoderm=360, Mechanism=First-m, Initialization=This paper, Pre-training corpus=Reddit2019.04 | — | 87.3 |