Molecular Property Optimization on QED maximize
74Normalized Rewardπ(·|y*)
Evaluation results
| Method | Links | |
|---|---|---|
| π(·|y*)2026.04 | 74 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=82026.04 | 62 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=162026.04 | 60 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=642026.04 | 60 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=322026.04 | 59 | |
| RLRL training budget=20002026.04 | 58 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=42026.04 | 54 | |
| RLRL training budget=10002026.04 | 52 | |
| RLRL training budget=5002026.04 | 43 | |
| Reward Weighted Classifier-Free Guidance (RCFG)Inference-time sampling budget |YS|=22026.04 | 40 |