Research map / Optimization / Implicit bias and generalization
Whitening Convergence Rate of Coupling-based Normalizing Flows
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 |
|---|---|---|---|
| Add coupling blocks with random orthogonal rotations | Coupling-based normalizing flows · Input data covariance | non-Standardness | Helped |
| Use affine-linear coupling layers to whiten the covariance | Coupling-based normalizing flows · Input data covariance | non-Standardness | Helped |
| Train a deeper Glow flow | Glow · EMNIST digits | non-Standardness | Helped |
| Add one coupling block while leaving previous blocks unchanged | Coupling-based normalizing flows · Toy covariance experiments | non-Standardness: approximately 50% | Helped |
| Pre-whiten data before passing it to the flow | Coupling-based normalizing flows · Input data covariance | non-Standardness | No clear effect |
Closest papers by meaning
- On the Universality of Volume-Preserving and Coupling-Based Normalizing Flows (ICML 2024 · similarity 0.89)
- Densely connected normalizing flows (NeurIPS 2021 · similarity 0.88)
- Fast and unified path gradient estimators for normalizing flows (ICLR 2024 · similarity 0.86)
- An Error Analysis of Flow Matching for Deep Generative Modeling (ICML 2025 · similarity 0.84)
- Flow matching achieves almost minimax optimal convergence (ICLR 2025 · similarity 0.84)
- Lifting Architectural Constraints of Injective Flows (ICLR 2024 · similarity 0.84)
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.