Research map / Optimization / Optimizers
Provable Guarantees for Gradient-Based 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 |
|---|---|---|---|
| Use FMRL instead of initializing from the previous task's optimal parameter | Mini-Wiki · Mini-Wiki | task-averaged regret | Helped |
| Use FLI-Batch approximate meta-updates instead of exact FAL updates (ablation) | FMRL · Mini-Wiki | task-averaged regret | Mixed |
| Use FMRL instead of first-order MAML | Mini-Wiki statistical setting · Mini-Wiki | meta-test accuracy | Mixed |
| Use the average iterate rather than the last iterate (ablation) | FMRL · Mini-Wiki batch setting | meta-test accuracy | No clear effect |
| Increase the number of training samples per task (ablation) | Reptile · Mini-ImageNet 5-shot 5-way; Omniglot 1-shot 5-way; Omniglot 5-shot 20-way | meta-test performance | Helped |
| Increase the number of within-task training iterations (ablation) | Reptile · Omniglot 5-way and 20-way; Mini-ImageNet 5-way | meta-test performance | Mixed |
Closest papers by meaning
- Adaptive Gradient-Based Meta-Learning Methods (NeurIPS 2019 · similarity 0.94)
- A Sample Complexity Separation between Non-Convex and Convex Meta-Learning (ICML 2020 · similarity 0.91)
- On the Stability and Generalization of Meta-Learning (NeurIPS 2024 · similarity 0.90)
- Optimistic Meta-Gradients (NeurIPS 2023 · similarity 0.89)
- The Advantage of Conditional Meta-Learning for Biased Regularization and Fine Tuning (NeurIPS 2020 · similarity 0.89)
- Memory-Reduced Meta-Learning with Guaranteed Convergence (AAAI 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.