Research map / Probabilistic and scientific modeling / Variational and generative sampling
Block Belief Propagation for Parameter Learning in Markov Random Fields
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 block-coordinate belief propagation with approximate gradient updates instead of full-network inference each learning iteration | MRF parameter learning · Synthetic grid networks, 200 × 200 | training speed: about three times faster than the inner-dual method; seven times faster than full convex BP | Helped |
| Use block-coordinate belief propagation with approximate gradient updates instead of full-network inference each learning iteration | MRF parameter learning · Synthetic BA networks, 20,000 nodes | training speed: about two times faster than inner-dual learning; three times faster than full convex BP | Helped |
| Use BBPL rather than the two full-network inference baselines | MRF training · Scene Understanding | objective reduction | Helped |
| Choose blocks about 20 to 25 times smaller than the network | BBPL · Synthetic grid and BA networks | convergence speed: 20 to 25 times smaller | Helped |
| Make blocks too small | BBPL · Synthetic grid and BA networks | iterations to converge | Hurt |
| Make blocks too large | BBPL · Synthetic grid and BA networks | per-iteration complexity | Hurt |
| Select subgraphs sequentially rather than randomly | BBPL · Synthetic networks | convergence speed | Helped |
Closest papers by meaning
- The Linearization of Belief Propagation on Pairwise Markov Random Fields (AAAI 2017 · similarity 0.89)
- Constraints Based Convex Belief Propagation (NeurIPS 2016 · similarity 0.86)
- Marginal Inference in Continuous Markov Random Fields Using Mixtures (AAAI 2019 · similarity 0.86)
- Inference in Graphical Models via Semidefinite Programming Hierarchies (NeurIPS 2017 · similarity 0.85)
- Belief Propagation Converges to Gaussian Distributions in Sparsely-Connected Factor Graphs (ICML 2026 · similarity 0.85)
- Learning in Deep Factor Graphs with Gaussian Belief Propagation (ICML 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.