Research map / Learning theory and algorithms / Kernels and regression theory
A PAC-Bayesian Generalization Bound for Equivariant Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Increase the equivariance group size |H| | H-equivariant models · G-MNIST | generalization error | Helped |
| Increase the equivariance group size |H| | H-equivariant models · Synthetic O2, F=6 | generalization error | Helped |
| Increase the equivariance group size |H| | H-equivariant models · Synthetic O2, F=1 | generalization error | No clear effect |
| Use Fourier-domain perturbation analysis rather than the naive PAC-Bayesian approach (ablation) | Equivariant networks · Synthetic O2 and SO2 datasets | correlation with generalization error | Helped |
| Use the architecture-only term Σ_l √M(l, η) as a predictor | H-equivariant models · Synthetic O2, F=6 | correlation with generalization error | Helped |
| Use the architecture-only term Σ_l √M(l, η) as a predictor | H-equivariant models · Synthetic O2, F=1 | correlation with generalization error | Hurt |
Closest papers by meaning
- Improving Equivariant Networks with Probabilistic Symmetry Breaking (ICLR 2025 · similarity 0.89)
- On the Generalization of Equivariant Graph Neural Networks (ICML 2024 · similarity 0.88)
- A Tale of Two Symmetries: Exploring the Loss Landscape of Equivariant Models (NeurIPS 2025 · similarity 0.88)
- A Canonicalization Perspective on Invariant and Equivariant Learning (NeurIPS 2024 · similarity 0.88)
- Generalization Bounds via Meta-Learned Model Representations: PAC-Bayes and Sample Compression Hypernetworks (ICML 2025 · similarity 0.88)
- Learning (Approximately) Equivariant Networks via Constrained Optimization (NeurIPS 2025 · similarity 0.87)
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.