Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
MAML is a Noisy Contrastive Learner in Classification
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 |
|---|---|---|---|
| Apply the zeroing trick during training (ablation) | FOMAML · mini-ImageNet | testing accuracy | Helped |
| Apply the zeroing trick during training (ablation) | SOMAML · mini-ImageNet | testing accuracy | Helped |
| Zero the linear layer at the beginning of meta-testing (ablation) | FOMAML · mini-ImageNet meta-testing | testing accuracy | Helped |
| Zero-initialize the linear layer (ablation) | FOMAML · mini-ImageNet | testing accuracy | Helped |
| Scale the initial linear layer by 0.7 (ablation) | FOMAML · mini-ImageNet | testing accuracy | Helped |
| Scale the initial linear layer by 0.5 (ablation) | FOMAML · mini-ImageNet | testing accuracy | Helped |
| Use one inner-loop update instead of more updates (ablation) | FOMAML with zeroing trick · mini-ImageNet 5-way 1-shot | convergence speed | No clear effect |
| Drop the interference term (ablation) | FOMAML reformulated loss · mini-ImageNet 5-way 1-shot | testing accuracy | Helped |
| Drop the noisy part of the contrastive coefficient (ablation) | FOMAML reformulated loss · mini-ImageNet 5-way 1-shot | testing accuracy | Helped |
| Apply the zeroing trick (ablation) | MAML · mini-ImageNet 5-way 1-shot | negative contrastive-term ratio: larger than 0.2 -> 0.2 | Helped |
| Apply the zeroing trick (ablation) | FOMAML · mini-ImageNet feature space | between-class cosine-similarity contrast | Helped |
| Use the zeroing trick (ablation) | FOMAML · mini-ImageNet, semantically similar classes | between-class cosine-similarity contrast | Helped |
Closest papers by meaning
- Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML (ICLR 2020 · similarity 0.91)
- Sharp-MAML: Sharpness-Aware Model-Agnostic Meta Learning (ICML 2022 · similarity 0.91)
- How to Train Your MAML to Excel in Few-Shot Classification (ICLR 2022 · similarity 0.90)
- How to train your MAML (ICLR 2019 · similarity 0.89)
- Generalization of Model-Agnostic Meta-Learning Algorithms: Recurring and Unseen Tasks (NeurIPS 2021 · similarity 0.88)
- Learning to Learn with Contrastive Meta-Objective (NeurIPS 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.