Agent Preference Modeling on Observed Preferences Full Fit
0.0921Modelling LossPer-Agent Per-Goal (Lower Bound)
Evaluation results
| Method | Links | |
|---|---|---|
| Per-Agent Per-Goal (Lower Bound)Hyper-parameters=7,1522026.05 | 0.0921 | |
| Per-Agent Per-Feature (Lower Bound)Hyper-parameters=2,9802026.05 | 0.1247 | |
| Proposed ModelHyper-parameters=572026.05 | 0.1513 | |
| SimultaneousHyper-parameters=572026.05 | 0.1534 | |
| QuadraticHyper-parameters=672026.05 | 0.1605 | |
| Diagonal SHyper-parameters=122026.05 | 0.1765 | |
| MemorylessHyper-parameters=572026.05 | 0.2041 | |
| Uniform Random (Baseline)Hyper-parameters=02026.05 | 0.3679 |