Research map / Graph learning / Graph algorithms and clustering
End to end learning and optimization on graphs
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 a differentiable clustering proxy with decision-focused training instead of standard two-stage learning or direct end-to-end prediction | ClusterNet · Community detection on Cora, Citeseer, Protein, Adol, and Facebook | modularity: Cora: 0.20 -> 0.54; Citeseer: 0.42 -> 0.55; Protein: 0.18 -> 0.29; Adol: 0.31 -> 0.49; Facebook: 0.25 -> 0.30 | Helped |
| Use a differentiable clustering proxy with decision-focused training instead of two-stage learning or direct end-to-end prediction | ClusterNet · Facility location on Cora, Citeseer, Protein, Adol, and Facebook | minmax distance: Cora: 12 -> 10; Citeseer: 17 -> 14; Protein: 8 -> 6; Adol: 7 -> 6; Facebook: 6 -> 4 | Helped |
| Train a graph reconstruction model before optimization rather than optimizing only on observed edges | Two-stage approaches · Community detection on single graphs | modularity | Mixed |
| Fine-tune ClusterNet on the observed edges of each test graph | ClusterNet · Inductive community detection, synthetic graphs | Avg.: 0.57 -> 0.60 | Helped |
| Fine-tune ClusterNet on the observed edges of each test graph | ClusterNet · Inductive community detection, PubMed | Avg.: 0.30 -> 0.40 | Helped |
| Train ClusterNet using one training graph instead of the full training set | ClusterNet · Inductive community detection, synthetic graphs | Avg.: 0.57 -> 0.55 | No clear effect |
| Train ClusterNet using one training graph instead of the full training set | ClusterNet · Inductive community detection, PubMed | Avg.: 0.30 -> 0.25 | Hurt |
Closest papers by meaning
- Learning Combinatorial Optimization Algorithms over Graphs (NeurIPS 2017 · similarity 0.88)
- Combinatorial Optimization with Graph Convolutional Networks and Guided Tree Search (NeurIPS 2018 · similarity 0.88)
- COMBHelper: A Neural Approach to Reduce Search Space for Graph Combinatorial Problems (AAAI 2024 · similarity 0.85)
- Towards General Algorithm Discovery for Combinatorial Optimization: Learning Symbolic Branching Policy from Bipartite Graph (ICML 2024 · similarity 0.85)
- Learning Graph Representation via Graph Entropy Maximization (ICML 2024 · similarity 0.84)
- Are Graph Neural Networks Optimal Approximation Algorithms? (NeurIPS 2024 · similarity 0.84)
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.