Research map / Architectures and efficiency / RNNs and normalization
Asynchronous Parallel Coordinate Minimization for MAP Inference
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 |
|---|---|---|---|
| adaptively reduce number of parallel workers based on estimated delay term (Proposition 1 criterion) (ablation) | APCM-Pencil · Synthetic 3-node chain graph | convergence of objective | Helped |
| increase workers from 1 to 40 with stale beliefs (ablation) | APCM-Pencil · Synthetic 3-node chain graph | convergence of objective | Hurt |
| use exact block coordinate minimization (Star updates) instead of stochastic gradient steps | APCM-Star vs HOGWILD! · Disparity estimation grid MRF | convergence speed of dual objective | Helped |
| asynchronous parallel coordinate minimization instead of HOGWILD! asynchronous SGD | MAP inference on disparity MRF · Disparity estimation grid MRF, connection widths 2/4/8 | speedup over HOGWILD! at same thread count: ~45x or higher | Helped |
| increase thread count from 1 to 46 for APCM-Star | APCM-Star · Disparity estimation grid MRF | speedup in threads | Helped |
Closest papers by meaning
- MAP Inference via Block-Coordinate Frank-Wolfe Algorithm (CVPR 2019 · similarity 0.87)
- Graph Cuts Always Find a Global Optimum for Potts Models (With a Catch) (ICML 2021 · similarity 0.87)
- Convergence of Some Convex Message Passing Algorithms to a Fixed Point (ICML 2024 · similarity 0.87)
- One-Pass Algorithms for MAP Inference of Nonsymmetric Determinantal Point Processes (ICML 2022 · similarity 0.86)
- Parallel AND/OR Search for Marginal MAP (AAAI 2020 · similarity 0.86)
- MAP Estimation with Denoisers: Convergence Rates and Guarantees (NeurIPS 2025 · 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.