5 problems
- 0 votes0 replies0 views
Conjecture on extending benign overfitting theory to super-wide deep neural networks
StoNet conjecture. The StoNet could enable the extension of benign overfitting theory from linear models to super-wide DNNs by leveraging its capability in sufficient dimension red…
- 0 votes0 replies0 views
Vanishing prior mass for over-dimensional subspaces in hierarchical Gaussian process estimation
Let be the central subspace and let the hierarchical prior assign mass to subspaces of varying dimension, including dimensions greater than the intrinsic dimension…
- 0 votes0 replies0 views
Exact recovery of the central subspace under hierarchical Gaussian process priors
Let and be fixed, and let denote the central subspace. Consider the hierarchical Bayes procedure described in the paper, whose poste…
- 0 votes0 replies0 views
Conjecture on oracle and selected-variable estimation matrices in sufficient dimension reduction
Let be the matrix formed using the first variables for the exact oracle sparse dimension-reduction estimator, and let be the correspondi…
- 0 votes0 replies0 views
Nonexistence of adaptive estimation without the linearity condition
In a single-index regression, the linearity condition is the condition under which the conditional mean of the covariates given the index is linear in the index, and an adaptive es…