Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Distribution Consistency Based Covariance Metric 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 |
|---|---|---|---|
| covariance (second-order) metric instead of GNN | few-shot classifier · Stanford Dogs 5-way 1-shot | accuracy: 46.98±0.98 -> 49.10±0.76 | Helped |
| covariance (second-order) metric instead of GNN | few-shot classifier · Stanford Dogs 5-way 5-shot | accuracy: 62.27±0.95 -> 63.04±0.65 | Helped |
| covariance (second-order) metric instead of GNN | few-shot classifier · Stanford Cars 5-way 1-shot | accuracy: 55.85±0.97 -> 56.65±0.86 | Helped |
| covariance (second-order) metric instead of GNN | few-shot classifier · Stanford Cars 5-way 5-shot | accuracy: 71.25±0.89 -> 71.33±0.62 | No clear effect |
| covariance (second-order) metric instead of GNN | few-shot classifier · CUB Birds 5-way 1-shot | accuracy: 51.83±0.98 -> 52.42±0.76 | Helped |
| covariance (second-order) metric instead of GNN | few-shot classifier · CUB Birds 5-way 5-shot | accuracy: 63.69±0.94 -> 63.76±0.64 | No clear effect |
| second-order covariance representation and metric instead of mean prototype | metric-based few-shot learner · miniImageNet 5-way 1-shot | accuracy: 50.44 -> 51.19 | Helped |
| CovaMNet vs Prototypical Nets (20-way trained) | few-shot classifier · miniImageNet 5-way 5-shot | accuracy: 68.20 -> 67.65 | Hurt |
| episodic training instead of pretrained network + k-NN | Conv-4 64F · miniImageNet 5-way 1-shot | accuracy: 27.23 -> 51.19 | Helped |
Closest papers by meaning
- Joint Distribution Matters: Deep Brownian Distance Covariance for Few-Shot Classification (CVPR 2022 · similarity 0.89)
- Improved Few-Shot Visual Classification (CVPR 2020 · similarity 0.87)
- Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning (CVPR 2021 · similarity 0.86)
- TADAM: Task dependent adaptive metric for improved few-shot learning (NeurIPS 2018 · similarity 0.86)
- Feature Distribution Fitting with Direction-Driven Weighting for Few-Shot Images Classification (AAAI 2023 · similarity 0.86)
- Adversarially Robust Few-shot Learning via Parameter Co-distillation of Similarity and Class Concept Learners (CVPR 2024 · similarity 0.85)
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.