Research map / Graph learning / Graph algorithms and clustering
SimGFM: Simplifying Discrete Flow Matching for Graph Generation
Chunyu Luo, Yuankai Luo, Xiao-Ming Wu, Lei Shi
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 |
|---|---|---|---|
| reduce sampling steps from 200 to 10 (ablation) | SimGFM · QM9 (no explicit H) | Valid (%): 99.8 -> 99.5 | No clear effect |
| reduce sampling steps from 200 to 10 (ablation) | SimGFM · QM9 (no explicit H) | FCD: 0.15 -> 0.92 | Hurt |
| reduce sampling steps from 200 to 10 (ablation) | SimGFM · QM9 (with explicit H) | Valid (%): 98.4 -> 93.7 | Hurt |
| reduce sampling steps from 200 to 10 (ablation) | SimGFM · MOSES | Validity (%): 89.4 -> 79.3 | Hurt |
| increase sampling steps from 50 to 1000 (ablation) | SimGFM · Planar | Ratio: 1.8 -> 21.0 | Hurt |
| increase sampling steps from 50 to 1000 (ablation) | SimGFM · Planar | V.U.N. (%): 99.5 -> 100.0 | Improved |
| use vf-denoiser rate with endpoint-focused polynomial scheduler instead of Campbell construction | graph DFM · QM9 | validity vs steps | Improved |
| add Endpoint Safety Projection (ESP) (ablation) | SimGFM · several benchmarks (appendix Table 10) | performance | Improved |
| use polynomial endpoint-focused scheduler instead of identity scheduler (k=1) (ablation) | SimGFM · appendix Table 11 | performance | Improved |
| use polynomial scheduler with Campbell's construction | Campbell rate matrix · graph generation | endpoint refinement | Hurt |
| add heuristic velocity terms (DeFoG) to Campbell field | DeFoG · some benchmarks | accuracy | No clear effect |
Closest papers by meaning
- Variational Flow Matching for Graph Generation (NeurIPS 2024 · similarity 0.91)
- DeFoG: Discrete Flow Matching for Graph Generation (ICML 2025 · similarity 0.91)
- Flow matching achieves almost minimax optimal convergence (ICLR 2025 · similarity 0.88)
- Discrete Flow Matching (NeurIPS 2024 · similarity 0.88)
- Flow Matching with General Discrete Paths: A Kinetic-Optimal Perspective (ICLR 2025 · similarity 0.88)
- Discrete-state Continuous-time Diffusion for Graph Generation (NeurIPS 2024 · similarity 0.87)
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.
