Offline Reinforcement Learning for Dialogue Management on Cornell Movie
3.62ReturnMoE-VRL
Evaluation results
| Method | Links | |
|---|---|---|
| MoE-VRLEvaluation Approach=Model Based2023.02 | 3.62 | |
| FtLEEvaluation Approach=Model Based2023.02 | 3.51 | |
| SAIQLEvaluation Approach=Model Based2023.02 | 2.61 | |
| EXP 1*Evaluation Approach=Model Based2023.02 | 1.47 | |
| FtLEEvaluation Approach=Model Free2023.02 | -0.39 | |
| MoE-VRLEvaluation Approach=Model Free2023.02 | -0.58 | |
| EXP 1*Evaluation Approach=Model Free2023.02 | -1.32 | |
| SAIQLEvaluation Approach=Model Free2023.02 | -1.34 |