Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
LGM-Net: Learning to Generate Matching Networks for Few-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 |
|---|---|---|---|
| Add intertask normalization (ablation) | Matching networks · miniImageNet 5-way 1-shot | accuracy: 44.98$\pm$0.88% -> 57.42$\pm$0.68% | Helped |
| Add intertask normalization (ablation) | Matching networks · miniImageNet 5-way 5-shot | accuracy: 55.74$\pm$0.75% -> 59.65$\pm$0.83% | Helped |
| Add the task context encoder (ablation) | LGM-Net · miniImageNet 5-way 1-shot | accuracy: 57.91$\pm$0.41% -> 69.13$\pm$ 0.35% | Helped |
| Add the task context encoder (ablation) | LGM-Net · miniImageNet 5-way 5-shot | accuracy: 59.89$\pm$0.48% -> 71.18$\pm$0.68% | Helped |
| Add intertask normalization (ablation) | LGM-Net · miniImageNet 5-way 1-shot | accuracy: 65.42$\pm$0.65% -> 69.13$\pm$ 0.35% | Helped |
| Add intertask normalization (ablation) | LGM-Net · miniImageNet 5-way 5-shot | accuracy: 67.93$\pm$0.56% -> 71.18$\pm$0.68% | Helped |
| Add randomness by sampling weight generation from a reparameterized multivariate Gaussian (ablation) | LGM-Net · miniImageNet 5-way 1-shot | accuracy: 67.25$\pm$0.42% -> 69.13$\pm$ 0.35% | Helped |
| Add randomness by sampling weight generation from a reparameterized multivariate Gaussian (ablation) | LGM-Net · miniImageNet 5-way 5-shot | accuracy: 69.68$\pm$0.55% -> 71.18$\pm$0.68% | Helped |
| Add weight normalization (ablation) | LGM-Net · miniImageNet 5-way 1-shot | accuracy: 68.85$\pm$0.41% -> 69.13$\pm$ 0.35% | Helped |
| Add weight normalization (ablation) | LGM-Net · miniImageNet 5-way 5-shot | accuracy: 69.97$\pm$0.45% -> 71.18$\pm$0.68% | Helped |
| Use an ensemble of generated weight points with max voting (ablation) | LGM-Net · miniImageNet 5-way 1-shot | accuracy: 69.13$\pm$ 0.35% -> 69.15$\pm$ 0.33% | No clear effect |
| Use an ensemble of generated weight points with max voting (ablation) | LGM-Net · miniImageNet 5-way 5-shot | accuracy: 71.18$\pm$0.68% -> 71.18$\pm$0.64% | No clear effect |
Closest papers by meaning
- Meta-Transfer Learning for Few-Shot Learning (CVPR 2019 · similarity 0.89)
- Data-Centric Meta-Learning for Robust Few-Shot Generalization (CVPR 2026 · similarity 0.88)
- Optimization as a Model for Few-Shot Learning (ICLR 2017 · similarity 0.88)
- TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning (ICML 2019 · similarity 0.87)
- Learning to Forget for Meta-Learning (CVPR 2020 · similarity 0.87)
- Zero-Shot Task Adaptation with Relevant Feature Information (AAAI 2024 · similarity 0.87)
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.