Research map / Learning theory and algorithms / Sample complexity
Task-Robust 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 |
|---|---|---|---|
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Sinusoid regression meta-test, 5-shot | worst-task MSE: 3.89 -> 2.82 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Sinusoid regression meta-test, 10-shot | worst-task MSE: 2.57 -> 1.68 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Sinusoid regression meta-test, 5-shot | mean MSE: 1.02 -> 1.09 | Hurt |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Sinusoid regression meta-test, 5-shot | std dev of task MSE: 0.88 -> 0.43 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Omniglot (20,5) meta-test alphabets | mean accuracy (%): 80.0 -> 87.6 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Omniglot (20,5) meta-test alphabets | worst alphabet accuracy (%): 67.9 -> 78.2 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Omniglot (20,1) meta-training alphabets | worst alphabet accuracy (%): 33.9 -> 82.4 | Helped |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Omniglot (5,1) meta-test alphabets | mean accuracy (%): 93.5 -> 93.1 | No clear effect |
| minimize worst-case task loss (TR-MAML) instead of average loss (MAML) | MAML · Omniglot (20,1) meta-training alphabets | weighted mean accuracy (%): 99.2 -> 92.2 | Hurt |
Closest papers by meaning
- Generalization of Model-Agnostic Meta-Learning Algorithms: Recurring and Unseen Tasks (NeurIPS 2021 · similarity 0.91)
- Task Agnostic Meta-Learning for Few-Shot Learning (CVPR 2019 · similarity 0.90)
- Probabilistic Model-Agnostic Meta-Learning (NeurIPS 2018 · similarity 0.90)
- Theoretical Investigations and Practical Enhancements on Tail Task Risk Minimization in Meta Learning (NeurIPS 2024 · similarity 0.90)
- A Simple Yet Effective Strategy to Robustify the Meta Learning Paradigm (NeurIPS 2023 · similarity 0.90)
- On the Stability and Generalization of Meta-Learning (NeurIPS 2024 · 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.