2" Elongation Prediction on Steel Property Prediction
3.611MAE (%)TabPFN
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| TabPFNFine-tuning Protocol=m.a.f.t. (multitask adapter fine-tuning)2026.03 | 3.611 | 75.56 | 46.12 | 90.98 | 9.73 | |
| TabPFNFine-tuning Protocol=m.f.t. (multitask fine-tuning)2026.03 | 3.623 | 74.73 | 46.19 | 90.93 | 9.64 | |
| TabPFNFine-tuning Protocol=s.f.t. (single-task fine-tuning)2026.03 | 3.642 | 74.64 | 46.01 | 90.83 | 9.29 | |
| TabPFNFine-tuning Protocol=n.f.t. (no fine-tuning)2026.03 | 3.656 | 74.52 | 46 | 90.79 | 8.04 | |
| XGBoostAlgorithm=Gradient Boosting2026.03 | 3.724 | 73.86 | 44.34 | 90.29 | 6.68 | |
| STLLearning Strategy=Single Task Learning2026.03 | 3.762 | 73.85 | 43.3 | 90 | 0 | |
| SAINTArchitecture=Transformer-based2026.03 | 3.781 | 72.86 | 43.7 | 90.03 | -0.77 | |
| MMoELearning Strategy=Mixture-of-Experts2026.03 | 3.79 | 72.43 | 43.36 | 89.81 | 0.35 | |
| PLELearning Strategy=Progressive Layered Extraction2026.03 | 3.826 | 72.25 | 43.03 | 89.57 | 0.61 | |
| FT-TransformerArchitecture=Transformer-based2026.03 | 3.864 | 71.88 | 41.93 | 89.74 | -2.66 | |
| MultiTabLearning Strategy=Multi-Tabular Learning2026.03 | 3.915 | 71.25 | 42.17 | 89.26 | 0.21 | |
| STEMLearning Strategy=Shared-Task Enhanced Model2026.03 | 3.956 | 70.77 | 41.47 | 88.8 | -0.4 | |
| MTLLearning Strategy=Multi-Task Learning2026.03 | 3.987 | 69.81 | 41.29 | 88.89 | -2.91 |