Research map / Learning theory and algorithms / Kernels and regression theory
Reproducing Kernel Banach Space Models for Neural Networks with Application to Rademacher Complexity Analysis
Alistair Shilton, Sunil Gupta, Santu Rana, Svetha Venkatesh
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Builds exact reproducing kernel Banach space (RKBS) models of feedforward networks (including ResNets and Transformers) via Hermite transforms and uses them for Rademacher complexity analysis.
Problem
Theoretical analysis of neural networks typically relies on approximations; the paper seeks exact, un-approximated function-space models of arbitrary-width/depth feedforward networks to analyze generalization via Rademacher complexity.
Approach
Uses Hermite transform based reproducing kernel Banach space methods to represent feedforward networks of arbitrary width, depth and topology exactly, assuming finite energy activations and finite spectral-norm weights and biases, and applies this representation to derive Rademacher complexity bounds.
Closest papers by meaning
- Why High-rank Neural Networks Generalize?: An Algebraic Framework with RKHSs (ICLR 2026 · similarity 0.89)
- Data-dependent Sample Complexity of Deep Neural Networks via Lipschitz Augmentation (NeurIPS 2019 · similarity 0.87)
- On the Rademacher Complexity of Graph Neural Networks: Unifying Expressivity and Geometry (NeurIPS 2026 · similarity 0.87)
- Gradient Descent in Neural Networks as Sequential Learning in Reproducing Kernel Banach Space (ICML 2023 · similarity 0.86)
- Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel (NeurIPS 2025 · similarity 0.85)
- Characterizing ResNet's Universal Approximation Capability (ICML 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
