Manifold Regularization
Cross-source consensus on Manifold Regularization from 1 sources and 4 claims.
1 sources · 4 claims
How it works
Benefits
Risks & contraindications
Highlighted claims
- The manifold penalty is applied to pre-activations so that even features removed by TopK can receive gradients. — GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models
- GeoSAE builds a k-nearest-neighbor graph over layer representations using Gaussian kernel weights. — GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models
- Removing manifold regularization reduced alive features and lowered conversion-prediction AUC. — GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models
- Too much manifold regularization over-regularized the representation and degraded performance. — GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models