Research map / Architectures and efficiency / CNN design
Enabling Equivariance for Arbitrary Lie Groups
Check 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 |
|---|---|---|---|
| use affine group convolutions | Image classification trained on MNIST · affNIST | accuracy: ~30% improvement | Helped |
| use homography group convolutions | Image classification trained on MNIST · homNIST | robustness | Helped |
| decrease the number of convolution or pooling samples from N=100 to N=25 (ablation) | Proposed group-convolutional model · affNIST and homNIST | test accuracy | Hurt |
| use only one sample to approximate convolution or pooling (ablation) | Proposed group-convolutional model · affNIST and homNIST | test accuracy | Hurt |
| decrease convolution samples rather than pooling samples (ablation) | Proposed group-convolutional model · affNIST and homNIST | test accuracy | Hurt |
| vary the number of samples used to approximate higher-layer convolutions (ablation) | Proposed group-convolutional model · Group-convolution layers | equivariance error | No clear effect |
Closest papers by meaning
- Lie Group Decompositions for Equivariant Neural Networks (ICLR 2024 · similarity 0.92)
- Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data (ICML 2020 · similarity 0.90)
- Variational Partial Group Convolutions for Input-Aware Partial Equivariance of Rotations and Color-Shifts (ICML 2024 · similarity 0.89)
- Affine Equivariant Networks Based on Differential Invariants (CVPR 2024 · similarity 0.88)
- On Making $SE(2)$-Invariant Networks Optimal (NeurIPS 2026 · similarity 0.87)
- Group Downsampling with Equivariant Anti-aliasing (ICLR 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.