Configuration Selection on LODO Leave-One-Dataset-Out (avg across datasets)
59PerformanceBOHB
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BOHBBaseline category=Sequential Baselines2026.03 | 59 | 3.8 | 95.5 | |
| AutoPipe (Ours)Phase=Two-phase, Evaluation fidelity=ES (early-stop)2026.03 | 59 | 3.8 | 8.71 | |
| SMACEvaluation fidelity=Full, Baseline category=Sequential Baselines2026.03 | 58 | 4.8 | 64 | |
| TPEEvaluation fidelity=Full, Baseline category=Sequential Baselines2026.03 | 57 | 5.8 | 66.8 | |
| Random SearchEvaluation fidelity=Full, Baseline category=Sequential Baselines2026.03 | 56.5 | 6.3 | 58.5 | |
| Offline-only (Ranker)Baseline category=Static & Zero-Shot Baselines2026.03 | 55.8 | 7 | 6.16 | |
| TPEEvaluation fidelity=ES (early-stop), Baseline category=Sequential Baselines2026.03 | 54.5 | 8.3 | 6.27 | |
| GlobalBaseline category=Static & Zero-Shot Baselines2026.03 | 52.9 | 9.9 | 4.05 | |
| SMACEvaluation fidelity=ES (early-stop), Baseline category=Sequential Baselines2026.03 | 50.4 | 12.3 | 5.99 | |
| Random SearchEvaluation fidelity=ES (early-stop), Baseline category=Sequential Baselines2026.03 | 44.6 | 18.2 | 4.95 | |
| LLM Rec. (History-conditioned)Baseline category=Static & Zero-Shot Baselines2026.03 | 44.2 | 18.6 | 5.59 | |
| Random (Single run)Baseline category=Static & Zero-Shot Baselines2026.03 | 39.1 | 23.7 | 3.62 | |
| LLM Rec. (Description-only)Baseline category=Static & Zero-Shot Baselines2026.03 | 30.7 | 32.1 | 7.6 |