AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Parametric Dual Maximization for Non-Convex Learning Problems

Yuxun Zhou, Zhaoyi Kang, Costas Spanos

AAAI 2017 · Optimization › Nonconvex and smooth optimization · approach: Coordinate and Adam convergence analysis · 7 citations (OpenAlex)

Read the paper

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
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · 2moons S3VMtest error rate (%): 1.9 -> 0.2Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · coil S3VMtest error rate (%): 5.3 -> 1.1Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · robot S3VMtest error rate (%): 9.5 -> 3.3Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · Wave LSVMtest error rate (%): 19.9 -> 13.1Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · 2moons S3VMnormalized objective value: 1.03 -> 1.00Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · Wave LSVMnormalized objective value: 1.54 -> 1.00Improved
increase approximation degree m from 0.1 to 0.2 length(p) (ablation)PDM · Wave LSVMtime: 2.5h -> 4.8hHurt
increase m beyond about 300 (range 80 to 650) (ablation)PDM · robot S3VMobjective and testing accuracyNo clear effect
use PDM instead of local optimizers (GD, CCCP, AO, LCS)S3VM/LSVM training · 2moons S3VMtest error rate (%): 51.4 -> 1.9Improved
use PDM instead of local optimizersLSVM training · Wave LSVMtest error rate (%): 36.7 -> 13.1Improved
use PDM instead of local optimizersLSVM training · Wave LSVMnormalized objective value: 10.2 -> 1.54Improved
use PDM instead of exact branch and boundS3VM · robot S3VMfeasibilityImproved

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