Research map / Graph learning / Graph algorithms and clustering
NN-Baker: A Neural-network Infused Algorithmic Framework for Optimization Problems on Geometric Intersection Graphs
Evan McCarty, Qi Zhao, Anastasios Sidiropoulos, Yusu Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| wrap TGS in the Baker framework | TGS · 2D-dense | MIS ratio: 0.915 -> 0.936 | Improved |
| wrap LwD in the Baker framework | LwD · 2D-dense | MIS ratio: 0.917 -> 0.955 | Improved |
| wrap Erdos-GNN in the Baker framework | Erdos-GNN · 2D-dense | MIS ratio: 0.834 -> 0.923 | Improved |
| wrap Erdos-GNN in the Baker framework | Erdos-GNN · 2DGaussian | MIS ratio: 0.769 -> 0.848 | Improved |
| wrap LwD in the Baker framework | LwD · 3D | MIS ratio: 0.902 -> 0.954 | Improved |
| wrap TGS in the Baker framework | TGS · Torus-4D | MIS ratio: 0.923 -> 0.937 | Improved |
| wrap LwD in the Baker framework | LwD · 2D-sparse | MIS ratio: 0.908 -> 0.926 | Improved |
| use Erdos-Baker instead of UNet-Baker | NN-Baker · 2D-sparse | MIS ratio: 0.919 -> 0.917 | No clear effect |
| replace KaMIS per-cell solver with NN (UNet-Baker vs K-Baker at similar runtime) | K-Baker · 2D-dense | MIS ratio: 0.753 -> 0.915 | Improved |
| replace exact KaMIS by UNet-Baker | KaMIS · 2D-dense | solve time (s): 3385.1 -> 18.06 | Improved |
| replace UNet with multi-layer fully connected network as Baker component (ablation) | UNet-Baker · MIS | MIS ratio: 70%-80% | Hurt |
| increase graph size without Baker decomposition | TGS/LwD/Erdos-GNN · 2D-dense | MIS ratio | Hurt |
Closest papers by meaning
- Are Graph Neural Networks Optimal Approximation Algorithms? (NeurIPS 2024 · similarity 0.89)
- Geometric Algorithms for Neural Combinatorial Optimization with Constraints (NeurIPS 2025 · similarity 0.88)
- COMBHelper: A Neural Approach to Reduce Search Space for Graph Combinatorial Problems (AAAI 2024 · similarity 0.87)
- On the Power of Small-size Graph Neural Networks for Linear Programming (NeurIPS 2024 · similarity 0.86)
- BIPNN: Learning to Solve Binary Integer Programming via Hypergraph Neural Networks (NeurIPS 2025 · similarity 0.85)
- On the Rademacher Complexity of Graph Neural Networks: Unifying Expressivity and Geometry (NeurIPS 2026 · 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.
