Molecular Property Optimization on Metabolically Stable
96Normalized Rewardπ(·|y*)
Evaluation Results
| Method | Links | |
|---|---|---|
| π(·|y*)2026.04 | 96 | |
| RLRL training budget=20002026.04 | 95 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=642026.04 | 92 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=322026.04 | 89 | |
| RLRL training budget=10002026.04 | 88 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=162026.04 | 83 | |
| RLRL training budget=5002026.04 | 78 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=82026.04 | 77 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=42026.04 | 61 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=22026.04 | 35 |