SynthCore

Training ideas

Curricula, sampling schedules, pseudo-labeling and multi-task mixing, with the standard training recipe as the control.

3 ideas found

Status: Mechanism failed Category: Training Math: Linear algebra Clear all
✗ Mechanism failed 2026

Inverse-Laplacian Residual Loss

Replace the standard squared pointwise PDE residual in an elliptic PINN by its discrete $H^{-1}$ norm. The residual is passed through an inverse Dirichlet Laplacian, reducing the dominance of rapidly varying residual modes and acting as a mathematically specified preconditioner for the PINN training gradients.

Useful8/10
Difficulty5/10
Novelty7/10
Training StabilitySpeedupAccuracy
Paper: Mitigating Numerical Stiffness in Least-Squares Formulations of Elliptic PDEs for Physics-Informed Neural Networks arXiv:2607.02726
✗ Mechanism failed 2026

Tail-triggered adaptive ridge head

Replace a fixed ridge coefficient in a neural network's final head with a controller driven by inverse spectral mass and hard-edge mass. The head can remain weakly regularized when the feature spectrum is healthy, but automatically increases ridge strength when small eigenvalues signal a high-risk interpolation regime.

Useful6/10
Difficulty4/10
Novelty5/10
Training StabilityGeneralizationAccuracy
Paper: High-Dimensional Interpolators Can Be Fragile: Heavy Tails and High-Dimensional Large Deviations arXiv:2607.09547
✗ Mechanism failed 2026

Extreme-Subset Adversarial Dropout

Turn row dropout into an adversarial conditioning problem rather than independent Bernoulli noise. At each training step, search for a subset of surviving channels or measurements with unusually small least singular value, train the downstream network on that subset, and gradually increase the search strength so training directly exposes failure modes hidden by average-case dropout.

Useful6/10
Difficulty6/10
Novelty8/10
Training StabilityGeneralizationSample efficiency
Paper: Extreme least singular values of Gaussian row submatrices and a phase retrieval stability problem arXiv:2607.06249