Research map / Learning theory and algorithms / Matrix and tensor methods
Sketching Method for Large Scale Combinatorial Inference
Wei Sun, Junwei Lu, Han Liu
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 |
|---|---|---|---|
| use sketching with sampled nodes instead of full-graph max clique test | fast max clique test · fMRI story data | runtime (seconds): 2649 -> 299 | Improved |
| sample 625 of 823 regions | fast max clique test · fMRI story data | testing accuracy | No clear effect |
| use sketching test instead of NLL at d=500 | connectivity test · Simulation, d=500 | runtime: 3.5 hours -> less than 2 minutes | Improved |
| increase disconnect ratio gamma | Ours · Simulation n=400, d=100, rho=0.25 | computation time | Improved |
| replace Direct_clime with Ours | connectivity test · n=400, d=100, rho=0.25 | power: 0 -> 0.67 | Improved |
| replace NLL with Ours | connectivity test · n=400, d=200, rho=0.3 | power: 0 -> 0.93 | Improved |
| replace Direct_clime with Ours | connectivity test · n=400, d=100, rho=0.25 | type-I error: 0 -> 0.19 | Hurt |
| replace Direct_node with Ours | connectivity test · n=400, d=100, rho=0.25 | type-I error: 0.98 -> 0.19 | Improved |
| increase sample size from n=400 to n=800 | Ours · d=100, rho=0.25 | power: 0.67 -> 1 | Improved |
| increase dimension from d=100 to d=200 | Ours · n=400, rho=0.25 | power: 0.67 -> 0.07 | Hurt |
| increase signal level rho from 0.25 to 0.3 | Ours · n=400, d=200 | power: 0.07 -> 0.93 | Improved |
| increase sample size from n=400 to n=800 | Direct_clime · d=100, rho=0.25 | power: 0 -> 0.55 | Improved |
Closest papers by meaning
- Matrix Product Sketching via Coordinated Sampling (ICLR 2025 · similarity 0.84)
- Distribution Learning Meets Graph Structure Sampling (NeurIPS 2025 · similarity 0.84)
- Sketch Down the FLOPs: Towards Efficient Networks for Human Sketch (CVPR 2025 · similarity 0.84)
- On the Power of the Weisfeiler-Leman Test for Graph Motif Parameters (ICLR 2024 · similarity 0.83)
- A Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening (NeurIPS 2024 · similarity 0.83)
- Sketch-Augmented Features Improve Learning Long-Range Dependencies in Graph Neural Networks (NeurIPS 2025 · similarity 0.83)
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.
