AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

New Convex Relaxations for MRF Inference With Unknown Graphs

Zhenhua Wang, Tong Liu, Qinfeng Shi, M. Pawan Kumar, Jianhua Zhang

ICCV 2019 · Learning theory and algorithms › Combinatorial and search optimization · approach: Predict-then-optimize learning · 7 citations (OpenAlex)

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 changedWhereMeasuredEffect
estimate the MRF graph structure instead of fixed tree-structured graph (ablation)tree-structured MRF · CADF1-score: 70.7 -> 74.4Improved
estimate the MRF graph structure instead of fixed tree-structured graph (ablation)tree-structured MRF · TVHIF1-score: 47.9 -> 62.2Improved
estimate the graph instead of using a known fully connected graph with TRW-S (ablation)fully connected MRF · TVHIF1-score: 55.6 -> 62.2Improved
use QP+CCCP instead of LP-W+B&BLP-W+B&B · CADF1-score: 73.6 -> 74.4Improved
use QP+CCCP instead of LP-W+B&BLP-W+B&B · TVHIF1-score: 61.1 -> 62.2Improved
use QP+CCCP instead of LP-W+B&BLP-W+B&B · BITF1-score: 54.2 -> 54.4Improved
use new LP-M relaxation instead of LP-WLP-W · CADF1-score: 73.6 -> 73.9Improved
use convex QP relaxation instead of LP-WLP-W · CADF1-score: 73.6 -> 71.7Hurt
add CCCP iterations on top of the convex QPQP · CADF1-score: 71.7 -> 74.4Improved
add CCCP iterations on top of the convex QP (ablation)QP · TVHIF1-score: 59.4 -> 62.2Improved
use QP+CCCP or LP-M solver instead of Lan's inferenceLan [15] · CAD test setestimated energy: -9.62 -> -10.28Improved
use LP-M instead of QP+CCCP for speedQP+CCCP · CADrunning time: 2.0 × 10−1 -> 3.0 × 10−2Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy