Generalization on Synthetic Dataset
0.251R^2 (10%)AdverISF (Two-Stage)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AdverISF (Two-Stage)Training Mode=Two-Stage2026.02 | 0.251 | 0.253 | 0.42 | 0.565 | |
| MLP2026.02 | 0.201 | 0.23 | 0.421 | 0.468 | |
| AdverISF (Joint)Training Mode=Joint2026.02 | 0.179 | 0.254 | 0.423 | 0.355 | |
| infoR–LSF2026.02 | 0.165 | 0.147 | 0.406 | 0.435 | |
| VIB2026.02 | 0.159 | 0.147 | 0.393 | 0.478 |