REVIEW 13 cited by
A Note on the PAC Bayesian Theorem
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
We prove general exponential moment inequalities for averages of [0,1]-valued iid random variables and use them to tighten the PAC Bayesian Theorem. The logarithmic dependence on the sample count in the enumerator of the PAC Bayesian bound is halved.
Forward citations
Cited by 13 Pith papers
-
Beyond Modern Asymptotics for Log-Likelihood Ratios in Logistic Regression
The worst-case (1−δ) quantile of the logistic log-likelihood ratio is d log(en/d)+log(1/δ) for n≥d≥3, with d=2 at log log log n and d=1 at log(1/δ).
-
Statistically Meaningful Geometry (SMG) Beyond the Euclidean Paradigm, with Application to Generative AI
Connection-filtered horizontal learning on an Orlicz fiber bundle bounds OOD predictive variance by the diameter of the identifiable base and eliminates catastrophic forgetting by orthogonal projection of updates.
-
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.
-
Generalization of Gibbs and Langevin Monte Carlo Algorithms in the Interpolation Regime
New PAC-Bayes bounds for the Gibbs posterior remain non-vacuous in the interpolation regime and can be approximated by Langevin Monte Carlo, but the tight experimental numbers rely on an unproved random-label calibrat...
-
PAC-Bayesian Certificates for Quadratic Closed-Loop Control
PAC-Bayesian bounds are derived for quadratic closed-loop control via SLS parameterization, yielding Chernoff certificates for posteriors over responses, a mean-response deployment result, and a data-driven learning a...
-
Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
Introduces a fairness layer for deep learning models that guarantees output parity and an online primal-dual algorithm for aggregate fairness guarantees in streaming predictions with small batch sizes.
-
Bound to Disagree: Generalization Bounds via Certifiable Surrogates
A disagreement-based certificate bounds any model's true risk by adding a concentration bound on unlabeled-data disagreement to a certified surrogate's risk.
-
Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
A computable generalization bound for gradient flow via the loss path kernel is proposed, but the proof's key Rademacher complexity lemma is false due to an underestimated chaos term.
-
Generalization Guarantees for Multi-View Representation Learning and Application to Regularization via Gaussian Product Mixture Prior
Multi-view representation learning generalization is bounded by the MDL of latent variables, and a Gaussian product mixture regularizer built from these bounds improves test accuracy over VIB baselines.
-
Federated Learning with Nonvacuous Generalisation Bounds
Federated learning trains private local randomised predictors whose aggregation yields a global predictor with nonvacuous PAC-Bayesian generalisation bounds and near-centralized accuracy.
-
Sample Complexity and Decision-Theoretic Guarantees for Bayesian Model Averaging over Decision Trees with Catalan-Exponential Priors
Establishes a complete non-asymptotic theory of rational commitment thresholds for BMA over BDTs with Dirichlet-Multinomial leaves and Catalan-exponential tree-size priors.
-
Margin-Adaptive Confidence Ranking for Reliable LLM Judgement
Learning a margin-based confidence ranker for LLM judges improves agreement-target success in cascaded selective evaluation compared to heuristic confidence scores.
-
PAC-Bayes with Backprop
Training neural networks with PAC-Bayes objectives yields MNIST test error of 1.4% and a non-vacuous risk bound of 2.3%, much tighter than prior PAC-Bayes certificates.
Discussion (0). Continue with ORCID to comment.