Research map / Graph learning / GNN architectures
Topology-Aware Network Pruning using Multi-stage Graph Embedding and Reinforcement 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 |
|---|---|---|---|
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res110 · CIFAR-10 | Top-1 accuracy: 93.74 -> 94.31 | Helped |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res56 · CIFAR-10 | Top-1 accuracy: 94.66 -> 93.49 | Hurt |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res32 · CIFAR-10 | Top-1 accuracy: 92.08 -> 92.58 | Helped |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res20 · CIFAR-10 | Top-1 accuracy: 91.38 -> 91.31 | Hurt |
| Use topology-aware GNN-RL pruning instead of AGMC | ShuffleNet-v1 · CIFAR-100 | Top-1 accuracy: 65.26 -> 67.10 | Helped |
| Use topology-aware GNN-RL pruning instead of AGMC | ShuffleNet-v2 · CIFAR-100 | Top-1 accuracy: 66.28 -> 66.64 | Helped |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | VGG16 · ImageNet | Top-1 accuracy: 69.10 -> 70.99 | Helped |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res18 · ImageNet | Top-1 accuracy: 68.41 -> 68.66 | Helped |
| Use topology-aware GNN-RL pruning instead of the strongest listed baseline | Res50 · ImageNet | Top-1 accuracy: 74.20 -> 74.28 | Helped |
| Use topology-aware GNN-RL pruning instead of EagleEye | MobileNet-v1 · ImageNet, 30% FLOPs reduction | Top-1 accuracy: 70.90 -> 70.70 | Hurt |
| Use topology-aware GNN-RL pruning instead of AMC | MobileNet-v1 · ImageNet, 60% FLOPs reduction | Top-1 accuracy: 68.90 -> 69.50 | Helped |
| Use topology-aware GNN-RL pruning instead of AMC | MobileNet-v2 · ImageNet | Top-1 accuracy: 70.80 -> 70.04 | Hurt |
Closest papers by meaning
- N2N learning: Network to Network Compression via Policy Gradient Reinforcement Learning (ICLR 2018 · similarity 0.88)
- Network Pruning via Performance Maximization (CVPR 2021 · similarity 0.88)
- Serving Graph Compression for Graph Neural Networks (ICLR 2023 · similarity 0.88)
- AutoCompress: An Automatic DNN Structured Pruning Framework for Ultra-High Compression Rates (AAAI 2020 · similarity 0.86)
- Network Sparsity Unlocks the Scaling Potential of Deep Reinforcement Learning (ICML 2025 · similarity 0.86)
- Sparsest Models Elude Pruning: An Exposé of Pruning’s Current Capabilities (ICML 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.