Proposes pointwise Riemannian Dimension from feature eigenvalues to derive tighter, representation-aware generalization bounds for deep networks in the nonlinear regime.
Advances in Neural Information Processing Systems , volume=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.LG 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
Adaptivity in linear bandits for ε-best arm identification gives only logarithmic improvements on hypercube, ℓ2 ball, m-sets and multi-task settings but polynomial-factor gains on a specially constructed action set, enabled by an adaptive O(d log(1/δ)/ε²) ℓ2-norm estimator.
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
citing papers explorer
-
Pointwise Generalization in Deep Neural Networks
Proposes pointwise Riemannian Dimension from feature eigenvalues to derive tighter, representation-aware generalization bounds for deep networks in the nonlinear regime.
-
On the Power of Adaptivity for $\varepsilon$-Best Arm Identification in Linear Bandits
Adaptivity in linear bandits for ε-best arm identification gives only logarithmic improvements on hypercube, ℓ2 ball, m-sets and multi-task settings but polynomial-factor gains on a specially constructed action set, enabled by an adaptive O(d log(1/δ)/ε²) ℓ2-norm estimator.
-
Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.