Research map / Graph learning / GNN architectures
Sequential Graph Convolutional Network for Active 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 |
|---|---|---|---|
| Use CoreGCN for active sampling | ResNet-18 active learning · CIFAR-10, 10,000 labelled examples | accuracy | Helped |
| Use CoreGCN for active sampling | ResNet-18 active learning · CIFAR-100, 40% of training data | accuracy | Helped |
| Use CoreGCN for active sampling | DeepPrior active learning · ICVL | mean squared error | Helped |
| Use UncertainGCN for active sampling | ResNet-18 active learning · CIFAR-10 imbalanced, 10,000 labelled samples | mean average accuracy | Helped |
| Use CoreGCN for active sampling | ResNet-18 active learning · CIFAR-10 imbalanced | mean average accuracy | Helped |
| Replace the GCN sampler with a two-layer dense discriminator (ablation) | UncertainGCN on CIFAR-10 · CIFAR-10 | accuracy | Hurt |
| Use a one-layer rather than two-layer GCN (ablation) | UncertainGCN on CIFAR-10 · CIFAR-10 | accuracy | Hurt |
| Use a three-layer rather than two-layer GCN (ablation) | UncertainGCN on CIFAR-10 · CIFAR-10 | accuracy | Hurt |
| Increase dropout from 0.3 to 0.5 or 0.8 (ablation) | UncertainGCN on CIFAR-10 · CIFAR-10 | accuracy | Hurt |
| Increase hidden units to 256 or 512 (ablation) | UncertainGCN on CIFAR-10 · CIFAR-10 | accuracy | No clear effect |
| Decrease the uncertainty margin from 0.4 to 0.1 (ablation) | UncertainGCN · CIFAR-10 | accuracy | Helped |
| Decrease the uncertainty margin further below 0.1 (ablation) | UncertainGCN · CIFAR-10 | accuracy | No clear effect |
Closest papers by meaning
- L2-GCN: Layer-Wise and Learned Efficient Training of Graph Convolutional Networks (CVPR 2020 · similarity 0.91)
- GraphSAINT: Graph Sampling Based Inductive Learning Method (ICLR 2020 · similarity 0.90)
- Simplifying Graph Convolutional Networks (ICML 2019 · similarity 0.90)
- Deeper Insights Into Graph Convolutional Networks for Semi-Supervised Learning (AAAI 2018 · similarity 0.89)
- All in a Row: Compressed Convolution Networks for Graphs (ICML 2023 · similarity 0.89)
- GCNext: Towards the Unity of Graph Convolutions for Human Motion Prediction (AAAI 2024 · similarity 0.88)
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.