Research map / Vision-language and multimodal / CLIP and zero-shot
Matching Networks for One Shot Learning
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 |
|---|---|---|---|
| Train with episodic tasks that match one-shot inference | Matching Networks · ImageNet L_rand 5-way 1-shot | accuracy: 87.6% -> 93.2% | Helped |
| Use Matching Nets instead of a feature-matching baseline | Omniglot · Omniglot 20-way 1-shot | accuracy: 88.1% -> 93.8% | Helped |
| Add fully conditional embeddings (FCE) (ablation) | Matching Nets · miniImageNet 5-way 1-shot | accuracy: 41.2% -> 44.2% | Helped |
| Add fully conditional embeddings (FCE) (ablation) | Matching Nets with fine-tuning · miniImageNet 5-way 1-shot | accuracy: 42.4% -> 46.6% | Helped |
| Add fully conditional embeddings (FCE) (ablation) | Matching Nets · miniImageNet 5-way 5-shot | accuracy: 56.2% -> 57.0% | Helped |
| Add fully conditional embeddings (FCE) (ablation) | Matching Nets with fine-tuning · miniImageNet 5-way 5-shot | accuracy: 58.0% -> 60.0% | Helped |
| Fine-tune on the one-shot support set (ablation) | Matching Nets · Omniglot 5-way 1-shot | accuracy: 98.1% -> 97.9% | Hurt |
| Fine-tune on the one-shot support set (ablation) | Matching Nets · miniImageNet 5-way 1-shot | accuracy: 41.2% -> 42.4% | Helped |
| Fine-tune on the one-shot support set (ablation) | Matching Nets · miniImageNet 5-way 5-shot | accuracy: 56.2% -> 58.0% | Helped |
| Evaluate on fine-grained dog classes after training on randomly sampled class tasks | Matching Nets + FCE · ImageNet L_dogs 5-way 1-shot | accuracy: 59.8% -> 58.8% | Hurt |
| Increase support examples per class from k=1 to k=3 (ablation) | Matching Networks · Penn Treebank 5-way | accuracy: 32.4% -> 38.2% | Helped |
| Add fully conditional embeddings (FCE) (ablation) | Matching Networks · Omniglot | accuracy | No clear effect |
Closest papers by meaning
- Memory Matching Networks for One-Shot Image Recognition (CVPR 2018 · similarity 0.90)
- Learning feed-forward one-shot learners (NeurIPS 2016 · similarity 0.90)
- Multi-Attention Network for One Shot Learning (CVPR 2017 · similarity 0.89)
- Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning (CVPR 2025 · similarity 0.86)
- Simple Semantic-Aided Few-Shot Learning (CVPR 2024 · similarity 0.86)
- Dual Memory Networks: A Versatile Adaptation Approach for Vision-Language Models (CVPR 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.