Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Scalable Bayesian Meta-Learning through Generalized Implicit Gradients
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 iBaML instead of MAML at K=5 | Bayesian meta-learning on a 4-layer 32-channel CNN · 5-class 5-shot miniImageNet | nll: 0.967 ± 0.017 -> 0.965 ± 0.018 | No clear effect |
| Use iBaML instead of MAML at K=5 | Bayesian meta-learning on a 4-layer 32-channel CNN · 5-class 5-shot miniImageNet | accuracy: 63.1 ± 0.92% -> 63.2 ± 0.74 % | No clear effect |
| Use iBaML instead of ABML at K=5 | Bayesian meta-learning on a 4-layer 32-channel CNN · 5-class 5-shot miniImageNet | nll: 0.957 ± 0.016 -> 0.965 ± 0.018 | No clear effect |
| Use iBaML instead of ABML at K=5 | Bayesian meta-learning on a 4-layer 32-channel CNN · 5-class 5-shot miniImageNet | accuracy: 62.8 ± 0.74 % -> 63.2 ± 0.74 % | No clear effect |
| Increase inner-level gradient steps from K=5 to K=10 (ablation) | iBaML · 5-class 5-shot miniImageNet | nll: 0.965 ± 0.018 -> 0.947 ± 0.017 | Helped |
| Increase inner-level gradient steps from K=5 to K=10 (ablation) | iBaML · 5-class 5-shot miniImageNet | accuracy: 63.2 ± 0.74 % -> 64.0 ± 0.75 % | Helped |
| Increase inner-level gradient steps from K=10 to K=15 (ablation) | iBaML · 5-class 5-shot miniImageNet | nll: 0.947 ± 0.017 -> 0.943 ± 0.017 | Helped |
| Increase inner-level gradient steps from K=10 to K=15 (ablation) | iBaML · 5-class 5-shot miniImageNet | accuracy: 64.0 ± 0.75 % -> 64.0 ± 0.74 % | No clear effect |
Closest papers by meaning
- Gradient-EM Bayesian Meta-Learning (NeurIPS 2020 · similarity 0.90)
- Continuous-Time Meta-Learning with Forward Mode Differentiation (ICLR 2022 · similarity 0.89)
- Memory-Reduced Meta-Learning with Guaranteed Convergence (AAAI 2025 · similarity 0.88)
- A Stochastic Approach to Bi-Level Optimization for Hyperparameter Optimization and Meta Learning (AAAI 2025 · similarity 0.87)
- More Flexible PAC-Bayesian Meta-Learning by Learning Learning Algorithms (ICML 2024 · similarity 0.87)
- Scalable Meta-Learning via Mixed-Mode Differentiation (ICML 2025 · similarity 0.86)
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.