Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian 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 |
|---|---|---|---|
| Use BMAML instead of MAML or EMAML | Sinusoidal regression · Sinusoidal regression test tasks | MSE | Helped |
| Use BMAML instead of EMAML | 5-way 1-shot image classification · miniImageNet | prediction accuracy | Helped |
| Use BMAML rather than EMAML with fewer meta-training tasks | 5-way 1-shot image classification · miniImageNet, 10K meta-training tasks | prediction robustness | Helped |
| Use Bayesian fast adaptation instead of EMAML | Active learning · miniImageNet | classification performance | Helped |
| Use BMAML rather than EMAML | Reinforcement learning · MuJoCo locomotion tasks | average reward | Mixed |
| Use SVPG-Chaser instead of VPG-Reptile | Reinforcement learning · 2D Navigation | average reward | Helped |
| Increase the number of particles (ablation) | BMAML · Sinusoidal regression | MSE | Helped |
| Increase the number of particles from 5 to 10 (ablation) | BMAML and EMAML · miniImageNet | prediction accuracy | Hurt |
| Increase the number of particles (ablation) | BMAML · Reinforcement learning | average reward | Mixed |
| Reduce the number of meta-training tasks (ablation) | BMAML versus EMAML · miniImageNet | prediction robustness | Helped |
| Increase the number of acquired examples during active learning (ablation) | Bayesian fast adaptation versus EMAML · miniImageNet active learning | classification performance gap | Helped |
Closest papers by meaning
- Gradient-EM Bayesian Meta-Learning (NeurIPS 2020 · similarity 0.92)
- Probabilistic Model-Agnostic Meta-Learning (NeurIPS 2018 · similarity 0.92)
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks (ICML 2017 · similarity 0.91)
- Amortized Bayesian Meta-Learning (ICLR 2019 · similarity 0.90)
- Recasting Gradient-Based Meta-Learning as Hierarchical Bayes (ICLR 2018 · similarity 0.89)
- More Flexible PAC-Bayesian Meta-Learning by Learning Learning Algorithms (ICML 2024 · similarity 0.88)
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.