Research map / Probabilistic and scientific modeling / Variational and generative sampling
Learning to Learn Kernels with Variational Random Features
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 |
|---|---|---|---|
| Add LSTM context inference to MetaVRF (ablation) | MetaVRF · miniImageNet 5-way classification | accuracy: 51.3$\pm$0.8 -> 53.1$\pm$0.9 (1-shot); 66.1$\pm$0.7 -> 66.8$\pm$0.7 (5-shot) | Helped |
| Add LSTM context inference to MetaVRF (ablation) | MetaVRF · CIFAR-FS 5-way classification | accuracy: 61.1$\pm$0.7 -> 62.1$\pm$0.8 (1-shot); 74.3 $\pm$0.9 -> 76.0$\pm$0.8 (5-shot) | Helped |
| Add LSTM context inference to MetaVRF (ablation) | MetaVRF · Omniglot classification | accuracy: 99.6$\pm$0.2 -> 99.7$\pm$0.2 (5-way 1-shot); 99.6$\pm$0.2 -> 99.8$\pm$0.1 (5-way 5-shot); 97.0$\pm$0.3 -> 97.5$\pm$0.3 ( | Mixed |
| Use bidirectional rather than vanilla LSTM inference (ablation) | MetaVRF · miniImageNet and CIFAR-FS 5-way classification | accuracy: 53.1$\pm$0.9 -> 54.2$\pm$0.8; 66.8$\pm$0.7 -> 67.8$\pm$0.7; 62.1$\pm$0.8 -> 63.1$\pm$0.7; 76.0$\pm$0.8 -> 76.5$\pm$0.9 | Helped |
| Use bidirectional rather than vanilla LSTM inference (ablation) | MetaVRF · Omniglot classification | accuracy: 99.7$\pm$0.2 -> 99.8$\pm$0.1; 99.8$\pm$0.1 -> 99.9$\pm$0.1; 97.5$\pm$0.3 -> 97.8$\pm$0.3; 99.0$\pm$0.2 -> 99.2$\pm$0.2 | Mixed |
| Use adaptive learned random Fourier features instead of a fixed RBF kernel | few-shot kernel learner · miniImageNet, CIFAR-FS, and Omniglot | accuracy | Helped |
| Learn adaptive random features instead of using regular RFFs | kernel learner · few-shot classification | accuracy | Helped |
Closest papers by meaning
- MetaModulation: Learning Variational Feature Hierarchies for Few-Shot Learning with Fewer Tasks (ICML 2023 · similarity 0.89)
- Bayesian Meta-Learning for the Few-Shot Setting via Deep Kernels (NeurIPS 2020 · similarity 0.88)
- Stochastic Variational Deep Kernel Learning (NeurIPS 2016 · similarity 0.87)
- TRF: Learning Kernels with Tuned Random Features (AAAI 2022 · similarity 0.87)
- Kernel Continual Learning (ICML 2021 · similarity 0.87)
- Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts (ICML 2024 · similarity 0.85)
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.