Conversational Response Selection 벤치마크
Conversational Response Selection on RRS Ranking Test
NDCG@3
- 2019-04-22 — Poly-encoder: NDCG@3 0.679
| Rank | Model | NDCG@3 | NDCG@5 | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 1 | Poly-encoder | 0.679 | 0.765 | Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring | sfzhou5678/PolyEncoder · chijames/Poly-Encoder · csong27/collision-bert · +4 | 2019 |
| 2 | SA-BERT+BERT-FP | 0.674 | 0.753 | Speaker-Aware BERT for Multi-Turn Response Selection in Retrieval-Based Chatbots | JasonForJoy/BERT-for-Response-Selection · JasonForJoy/SA-BERT | 2020 |
| 3 | BERT | 0.625 | 0.714 | An Effective Domain Adaptive Post-Training Method for BERT in Response Selection | taesunwhang/BERT-ResSel | 2019 |
| 4 | BERT-FP | 0.609 | 0.709 | Fine-grained Post-training for Improving Retrieval-based Dialogue Systems | hanjanghoon/BERT_FP | 2021 |