Hardness-aware kernelized contrastive regression
Intervention1 paper1 finding
- In silico1
Outcomes
1Mean absolute error
downin silico1
1 study
Mean absolute error
downin silico1
Hardness-aware kernelized contrastive regression decreases mean absolute error in silico (MAE 3.84 on OpenBHB internal, 4.72 on external, 3.01 on ADNI).
In silico
Hardness-Aware Kernelized Contrastive Learning for Brain Age Prediction · Research Square · 28 Aug 2026 · Preprint
Latest
1Research SquareIn silicoPreprint