Counterfactual Policy Evaluation on GridWorld p = 0.9
346Avg Worst-Case Counterfactual V(s0)ICFMDP
Evaluation Results
| Method | Links | |
|---|---|---|
| ICFMDPSampled observed trajectories=1002025.02 | 346 | |
| Gumbel-maxSampled observed trajectories=1002025.02 | 304 |
| Method | Links | |
|---|---|---|
| ICFMDPSampled observed trajectories=1002025.02 | 346 | |
| Gumbel-maxSampled observed trajectories=1002025.02 | 304 |