Research map / Architectures and efficiency / RNNs and normalization
LieGG: Studying Learned Lie Group Generators
Artem Moskalev, Anna Sepliarskaia, Ivan Sosnovik, Arnold W.M. Smeulders
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 training set size (ablation) | MLP · O(5) synthetic regression | symmetry variance | Improved |
| increase number of parameters (ablation) | MLP · rotation-MNIST | symmetry variance | Improved |
| deeper instead of wider network at same parameter count (ablation) | MLP · rotation-MNIST | symmetry variance | Improved |
| increase depth beyond 3 layers (ablation) | MLP · rotation-MNIST | symmetry bias | Mixed |
| use deeper 6-layer network instead of shallow 2-layer (ablation) | MLP · rotation-MNIST | symmetry bias | Improved |
| pre-train sub-network then fine-tune or layer-wise train (ablation) | MLP-7 · rotation-MNIST | intermediate layer symmetry bias and variance | Improved |
| vary training regime (end-to-end, fine-tune, layer-wise) (ablation) | MLP-7 · rotation-MNIST | final-layer symmetry | No clear effect |
| lower symmetry variance and bias vs accuracy | MLP · rotation-MNIST | classification accuracy: -0.49 | Mixed |
Closest papers by meaning
- Learning Infinitesimal Generators of Continuous Symmetries from Data (NeurIPS 2024 · similarity 0.91)
- Lie Algebra Canonicalization: Equivariant Neural Operators under arbitrary Lie Groups (ICLR 2025 · similarity 0.89)
- Symmetry Induces Structure and Constraint of Learning (ICML 2024 · similarity 0.89)
- Lie Group Decompositions for Equivariant Neural Networks (ICLR 2024 · similarity 0.89)
- Explicit Discovery of Nonlinear Symmetries from Dynamic Data (ICML 2025 · similarity 0.89)
- Improving Equivariant Networks with Probabilistic Symmetry Breaking (ICLR 2025 · similarity 0.88)
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.
