Research map / Learning theory and algorithms / Combinatorial and search optimization
Min Norm Point Algorithm for Higher Order MRF-MAP Inference
Check 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 |
|---|---|---|---|
| Exploit the sum-of-submodular structure during Min Norm Point inference | Min Norm Point for synthetic edge-based costs · Clique size 2 × 2 | inference time (seconds): 0.776 -> 0.006 | Helped |
| Use SoS-MinNorm instead of GC | Synthetic edge-based costs, problem size 400 · Clique size 4 × 2 | inference time (seconds): 0.26 -> 0.01 | Helped |
| Use SoS-MinNorm instead of GC | Synthetic edge-based costs · Clique size 4 × 4 | inference time (seconds): 731.67 -> 0.02 | Helped |
| Use SoS-MinNorm instead of SoS-Jegelka | Synthetic edge-based costs · Clique size 4 × 2 | inference time (seconds): 12217.37 -> 0.01 | Helped |
| Scale the problem size while keeping clique size 16 | SoS-MinNorm · Synthetic edge-based costs, problem sizes 100, 400, 900, and 1600 | inference time (seconds): 0.01, 0.01, 0.04, 0.09 | Mixed |
| Use SoS-MinNorm instead of GC | Synthetic edge-based costs with clique size 16 · Problem size 1600 | inference time (seconds): 2868 -> 0.09 | Helped |
| Increase clique size from small to large | SoS-MinNorm on pixel-level object detection · Image size 100 × 100 | visual result quality | Helped |
| Increase clique size | SoS-MinNorm on interactive object segmentation · Image size 103 × 132 | visual result quality | Helped |
| Increase the stopping tolerance ǫ (ablation) | SoS-MinNorm approximate inference · Pixel-level object detection with clique size ∼250 | inference time (seconds): 0.03 -> 3.66 | Helped |
| Increase the stopping tolerance ǫ (ablation) | SoS-MinNorm approximate inference · Pixel-level object detection with clique size ∼250 | primal-dual gap | Hurt |
| Increase the stopping tolerance ǫ (ablation) | SoS-MinNorm approximate inference · Pixel-level object detection with clique size ∼250 | visual result quality | Hurt |
Closest papers by meaning
- Inference in Higher Order MRF-MAP Problems With Small and Large Cliques (CVPR 2018 · similarity 0.93)
- Continuous DR-submodular Maximization: Structure and Algorithms (NeurIPS 2017 · similarity 0.87)
- Efficient semidefinite-programming-based inference for binary and multi-class MRFs (NeurIPS 2020 · similarity 0.87)
- Continuous Relaxation of MAP Inference: A Nonconvex Perspective (CVPR 2018 · similarity 0.87)
- Inferring Neural Signed Distance Functions by Overfitting on Single Noisy Point Clouds through Finetuning Data-Driven based Priors (NeurIPS 2024 · similarity 0.84)
- Memory-Scalable and Simplified Functional Map Learning (CVPR 2024 · 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.