Research map / Label-efficient and robust learning / Domain adaptation and generalization
MetaReg: Towards Domain Generalization using Meta-Regularization
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 meta-learned weighted L1 regularizer | AlexNet · PACS | average accuracy: 69.28 -> 72.62 | Helped |
| add meta-learned regularizer | ResNet-18 · PACS | average accuracy: 79.9 -> 81.7 | Helped |
| add meta-learned regularizer | ResNet-50 · PACS Photo | accuracy: 97.8 -> 97.6 | No clear effect |
| add meta-learned regularizer | MLP · Amazon Reviews | average accuracy: 80.7 -> 81.2 | Helped |
| use weighted L1 instead of weighted L2 regularizer class (ablation) | MetaReg · PACS Art painting | accuracy: 83.2 -> 83.7 | Helped |
| learn regularizer weights instead of fixed L1 (phi=1) (ablation) | MetaReg · PACS Art painting | accuracy: 79.7 -> 83.7 | Helped |
| add DropConnect regularization (ablation) | ResNet-18 · PACS Art painting | accuracy: 79.9 -> 80.1 | No clear effect |
| train regularizer on more data (f 0.1 -> 1.0) (ablation) | MetaReg · PACS | accuracy: 82.86 -> 83.71 | No clear effect |
| regularize more task-network layers (ablation) | MetaReg AlexNet · PACS Cartoon | accuracy: 67.31 -> 70.35 | Helped |
Closest papers by meaning
- Learning to Generalize: Meta-Learning for Domain Generalization (AAAI 2018 · similarity 0.93)
- Open Domain Generalization with Domain-Augmented Meta-Learning (CVPR 2021 · similarity 0.91)
- Improving Generalization of Meta-Learning With Inverted Regularization at Inner-Level (CVPR 2023 · similarity 0.90)
- Compound Domain Generalization via Meta-Knowledge Encoding (CVPR 2022 · similarity 0.89)
- Improving Generalization via Meta-Learning on Hard Samples (CVPR 2024 · similarity 0.87)
- Seeking Consistent Flat Minima for Better Domain Generalization via Refining Loss Landscapes (CVPR 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.