Research map / Learning theory and algorithms / Combinatorial and search optimization
Constraints Based Convex Belief Propagation
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 |
|---|---|---|---|
| replace PN-consistency potentials with belief consistency constraints (CBCBP) | convex belief propagation · Synthetic MRF n=100-400 variables | running time: about 25x to 31x faster | Helped |
| increase PN-potential penalty c (speedup of constraints over PN potentials grows) | CBCBP vs CBP · Synthetic MRF n=200 variables | speedup factor: 31.40x at c=2 to 58.01x at c=10 | Helped |
| replace PN-consistency potentials with belief consistency constraints (CBCBP) | convex belief propagation · MSRC-21 | global accuracy: 84.2 to 85.4 | Helped |
| replace PN-consistency potentials with belief consistency constraints (CBCBP) | convex belief propagation · MSRC-21 | average accuracy: 74.3 to 76.1 | Helped |
| replace PN-consistency potentials with belief consistency constraints (CBCBP) | convex belief propagation · MSRC-21 | running time: 1.41s to 0.02s | Helped |
| replace PN-consistency potentials with belief consistency constraints (CBCBP) | convex belief propagation · PascalVOC 2012 | average pixel accuracy and IOU: 90.6 to 91.6 pixel accuracy; 62.69 to 62.97 IOU | Helped |
| add source word consistency constraints to phrase-based translation model (ablation) | CBCBP translation model · German-English WMT2009 first 150 sentences | %BLEU: 13.12 to 16.73 | Helped |
| replace PN-consistency potentials with belief consistency constraints (CBCBP) on the synthetic tasks of Table 1 and 2 in accuracy-free setting | convex belief propagation · MSRC-21 per-class accuracy | per-class accuracy (19 of 21 classes equal or higher): CBCBP lower on void (0.79 to 0.72) and some classes, higher on most | Mixed |
Closest papers by meaning
- The Linearization of Belief Propagation on Pairwise Markov Random Fields (AAAI 2017 · similarity 0.86)
- Min-Max Propagation (NeurIPS 2017 · similarity 0.86)
- Block Belief Propagation for Parameter Learning in Markov Random Fields (AAAI 2019 · similarity 0.86)
- Deep Attentive Belief Propagation: Integrating Reasoning and Learning for Solving Constraint Optimization Problems (NeurIPS 2022 · similarity 0.86)
- Convergence of Some Convex Message Passing Algorithms to a Fixed Point (ICML 2024 · similarity 0.85)
- The Theory and Practice of MAP Inference over Non-Convex Constraints (ICML 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.