Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
MetaKernel: Learning Variational Random Features With Limited Labels
Yingjun Du, Haoliang Sun, Xiantong Zhen, Jun Xu, Yilong Yin, Ling Shao, Cees G. M. Snoek
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 random Fourier features instead of RBF kernel (ablation) | MetaKernel · miniImageNet 5-way 1-shot | accuracy (%): 42.1 -> 52.8 | Improved |
| use random Fourier features instead of RBF kernel (ablation) | MetaKernel · CIFAR-FS 5-way 5-shot | accuracy (%): 59.8 -> 74.7 | Improved |
| use variational RFFs instead of plain RFFs (ablation) | MetaKernel · miniImageNet 5-way 1-shot | accuracy (%): 52.8 -> 51.3 | Hurt |
| use variational RFFs instead of plain RFFs (ablation) | MetaKernel · miniImageNet 5-way 5-shot | accuracy (%): 65.4 -> 66.1 | No clear effect |
| use variational RFFs instead of plain RFFs (ablation) | MetaKernel · CIFAR-FS 5-way 5-shot | accuracy (%): 74.7 -> 74.3 | No clear effect |
| add task context inference (LSTM) to variational RFFs (ablation) | MetaKernel · miniImageNet 5-way 1-shot | accuracy (%): 51.3 -> 54.2 | Improved |
| add task context inference (LSTM) to variational RFFs (ablation) | MetaKernel · CIFAR-FS 5-way 1-shot | accuracy (%): 61.1 -> 63.1 | Improved |
| add task context inference (LSTM) to variational RFFs (ablation) | MetaKernel · CIFAR-FS 5-way 5-shot | accuracy (%): 74.3 -> 76.5 | Improved |
| add conditional normalizing flows to enrich random features (ablation) | MetaKernel · miniImageNet 5-way 1-shot | accuracy (%): 54.2 -> 55.5 | Improved |
| add conditional normalizing flows to enrich random features (ablation) | MetaKernel · CIFAR-FS 5-way 1-shot | accuracy (%): 63.1 -> 64.3 | Improved |
| add conditional normalizing flows to enrich random features (ablation) | MetaKernel · miniImageNet 5-way 1-shot (pretrained WRN-28-10) | accuracy (%): 63.80 -> 65.03 | Improved |
| MetaKernel vs MetaNorm in few-shot domain generalization | MetaKernel · DomainNet 5-way | accuracy (%): 32.7 -> 34.7 (1-shot); 51.9 -> 53.7 (5-shot) | Improved |
Closest papers by meaning
- Learning to Learn Kernels with Variational Random Features (ICML 2020 · similarity 0.93)
- Bayesian Meta-Learning for the Few-Shot Setting via Deep Kernels (NeurIPS 2020 · similarity 0.89)
- Meta-Exploiting Frequency Prior for Cross-Domain Few-Shot Learning (NeurIPS 2024 · similarity 0.87)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.87)
- One Meta-tuned Transformer is What You Need for Few-shot Learning (ICML 2024 · similarity 0.87)
- Accelerating Convergence in Bayesian Few-Shot Classification (ICML 2024 · 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.
