Path Regularization: A Near-Complete and Optimal Nonasymptotic Generalization Theory for Multilayer Neural Networks and Double Descent Phenomenon
cs.LG, cs.AI, math.ST, stat.TH
Submitted: 2025-03-03
Updated: 2026-09-23
Terminology
Sources
- Gradient Descent Provably Optimizes Over-parameterized Neural Networks
- A path-norm toolkit for modern networks: consequences, promises and challenges
- Unraveling the Enigma of Double Descent: An In-depth Analysis through the Lens of Learned Feature Space
- Generalization in Deep Networks: The Role of Distance from Initialization
- A PAC-Bayesian Approach to Spectrally-Normalized Margin Bounds for Neural Networks
- Towards Understanding the Role of Over-Parametrization in Generalization of Neural Networks
- Double Descent Demystified: Identifying, Interpreting & Ablating the Sources of a Deep Learning Puzzle
- Deep Network Approximation Characterized by Number of Neurons
- Nonasymptotic theory for two-layer neural networks: Beyond the bias-variance trade-off
- Generalization Error Bounds for Deep Neural Networks Trained by SGD
- On the Banach spaces associated with multi-layer ReLU networks: Function representation, approximation theory and gradient descent dynamics
- Estimating the Generalization in Deep Neural Networks via Sparsity
- SGD Converges to Global Minimum in Deep Learning via Star-convex Path
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks