Research map / Optimization / Optimizers
Sharp-MAML: Sharpness-Aware Model-Agnostic Meta Learning
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 SAM to fine-tuning step only (Sharp-MAML_low) | MAML · Mini-Imagenet 5-way 1-shot | accuracy: +2.59 (47.13 to 49.72) | Helped |
| add SAM to fine-tuning and meta-update (Sharp-MAML_both) | MAML · Mini-Imagenet 5-way 5-shot | accuracy: +2.84 (62.20 to 65.04) | Helped |
| add SAM to meta-update step only (Sharp-MAML_up) | MAML · Mini-Imagenet 5-way 1-shot | accuracy: +2.43 (47.13 to 49.56) | Helped |
| add SAM to fine-tuning only (Sharp-MAML_low) | MAML · Mini-Imagenet 5-way 5-shot | accuracy: +0.98 (62.20 to 63.18) | Helped |
| replace SAM with ESAM in Sharp-MAML_low (ablation) | Sharp-MAML_low · Mini-Imagenet 5-way 1-shot | accuracy: -0.82 (49.72 to 48.90) | Hurt |
| replace fine-tuning with ANIL in Sharp-MAML_low (ablation) | Sharp-MAML_low · Mini-Imagenet 5-way 1-shot | accuracy: -0.53 (49.72 to 49.19) at about 86% less computation | Mixed |
Closest papers by meaning
- MAML is a Noisy Contrastive Learner in Classification (ICLR 2022 · similarity 0.91)
- How to train your MAML (ICLR 2019 · similarity 0.90)
- Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML (ICLR 2020 · similarity 0.90)
- Global Convergence of MAML and Theory-Inspired Neural Architecture Search for Few-Shot Learning (CVPR 2022 · similarity 0.90)
- On the Global Optimality of Model-Agnostic Meta-Learning (ICML 2020 · similarity 0.89)
- Repurposing Pretrained Models for Robust Out-of-domain Few-Shot Learning (ICLR 2021 · similarity 0.89)
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.