Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian Few-Shot Classification with One-vs-Each Pólya-Gamma Augmented Gaussian Processes
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 predictive likelihood instead of marginal likelihood (ablation) | OVE PG GP + Cosine · CUB 5-shot | accuracy (%): 77.44 -> 79.07 | Helped |
| Use predictive likelihood instead of marginal likelihood (ablation) | OVE PG GP + Cosine · mini-ImageNet 5-shot | accuracy (%): 64.58 -> 67.14 | Helped |
| Use predictive likelihood instead of marginal likelihood (ablation) | OVE PG GP + Cosine · Omniglot→EMNIST 1-shot | accuracy (%): 68.43 -> 77.00 | Helped |
| Use predictive likelihood instead of marginal likelihood (ablation) | OVE PG GP + Cosine · mini-ImageNet→CUB 5-shot | accuracy (%): 55.71 -> 57.23 | Helped |
| Replace cosine kernel with linear kernel (ablation) | OVE PG GP with marginal likelihood · CUB 1-shot | accuracy (%): 63.98 -> 62.48 | Hurt |
| Replace cosine kernel with linear kernel (ablation) | OVE PG GP with marginal likelihood · CUB 5-shot | accuracy (%): 77.44 -> 77.94 | Helped |
| Replace cosine kernel with RBF kernel (ablation) | OVE PG GP with predictive likelihood · CUB 1-shot | accuracy (%): 60.11 -> 56.18 | Hurt |
| Replace cosine kernel with RBF kernel (ablation) | OVE PG GP with predictive likelihood · Omniglot→EMNIST 5-shot | accuracy (%): 87.52 -> 89.87 | Helped |
| Replace cosine kernel with normalized RBF kernel (ablation) | OVE PG GP with predictive likelihood · mini-ImageNet→CUB 1-shot | accuracy (%): 37.49 -> 37.50 | No clear effect |
| Optimize predictive likelihood rather than marginal likelihood (ablation) | OVE PG GP + Cosine · CUB 1-shot | accuracy (%): 63.98 -> 60.11 | Hurt |
| Optimize predictive likelihood rather than marginal likelihood (ablation) | OVE PG GP + Cosine · mini-ImageNet→CUB 1-shot | accuracy (%): 39.66 -> 37.49 | Hurt |
| Use predictive likelihood rather than marginal likelihood (ablation) | Logistic Softmax GP + Cosine · CUB 5-shot | accuracy (%): 74.58 -> 78.14 | Helped |
Closest papers by meaning
- Accelerating Convergence in Bayesian Few-Shot Classification (ICML 2024 · similarity 0.89)
- Revisiting Logistic-softmax Likelihood in Bayesian Meta-Learning for Few-Shot Classification (NeurIPS 2023 · similarity 0.89)
- Bayesian Exploration of Pre-trained Models for Low-shot Image Classification (CVPR 2024 · similarity 0.86)
- Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learning (CVPR 2024 · similarity 0.85)
- MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlation (ICLR 2024 · similarity 0.85)
- Frozen Feature Augmentation for Few-Shot Image Classification (CVPR 2024 · similarity 0.84)
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.