Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Instance Credibility Inference for Few-Shot Learning
Check 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 ICI-based pseudo-label selection to LR | LR · miniImageNet 1-shot, transductive | accuracy: 56.06 -> 66.80 | Helped |
| Add ICI-based pseudo-label selection to LR | LR · tieredImageNet 1-shot, transductive | accuracy: 69.02 -> 80.79 | Helped |
| Add ICI-based pseudo-label selection to LR | LR · CIFAR-FS 1-shot, transductive | accuracy: 62.25 -> 73.97 | Helped |
| Add ICI-based pseudo-label selection to LR | LR · CUB 1-shot, transductive | accuracy: 76.16 -> 88.06 | Helped |
| Increase available unlabeled samples to 80 per class | LR + ICI · miniImageNet 1-shot, semi-supervised | accuracy: 65.86 -> 71.41 | Helped |
| Increase available unlabeled samples to 80 per class | LR + ICI · tieredImageNet 1-shot, semi-supervised | accuracy: 81.10 -> 85.44 | Helped |
| Replace ICI with random unlabeled-sample selection (ablation) | Self-taught LR pipeline · miniImageNet 1-shot, transductive | accuracy: 65.32 -> 59.01 | Hurt |
| Replace ICI with nearest-neighbor selection in feature space (ablation) | Self-taught LR pipeline · miniImageNet 1-shot, transductive | accuracy: 65.32 -> 63.24 | Hurt |
| Replace ICI with classifier-confidence selection (ablation) | Self-taught LR pipeline · miniImageNet 1-shot, transductive | accuracy: 65.32 -> 63.29 | Hurt |
| Select samples iteratively rather than all at once (ablation) | ICI · miniImageNet transductive | accuracy | Helped |
| Use PCA feature dimension 5 rather than 2 (ablation) | ICI · miniImageNet 5-way 1-shot | accuracy: 63.71±1.025 -> 66.80±1.096 | Helped |
| Use PCA feature dimension 5 rather than 10 (ablation) | ICI · miniImageNet 5-way 1-shot | accuracy: 66.25±1.048 -> 66.80±1.096 | Helped |
Closest papers by meaning
- Few-Shot Learning With Noisy Labels (CVPR 2022 · similarity 0.88)
- An Embarrassingly Simple Approach to Semi-Supervised Few-Shot Learning (NeurIPS 2022 · similarity 0.88)
- Meta-Learning for Semi-Supervised Few-Shot Classification (ICLR 2018 · similarity 0.87)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.86)
- Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning (CVPR 2025 · similarity 0.86)
- MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlation (ICLR 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.