AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Projection onto Minkowski Sums with Application to Constrained Learning

Joong-Ho Won, Jason Xu, Kenneth Lange

ICML 2019 · Optimization › Nonconvex and smooth optimization · approach: Coordinate and Adam convergence analysis · 8 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
Minkowski projection block descent (Algorithm 1) used for overlapping group lasso proximal mapsoverlapping group lasso (l1,p) with p=2 · synthetic, d from 10^3 to 10^6, g=10-100runtime vs SLEPImproved
Minkowski projection method vs SLEP in high group count regimeoverlapping group lasso · synthetic, d=10^6runtimeHurt
Minkowski projection method vs SLEP in accuracyoverlapping group lasso · synthetic, varying dimension and groupsfinal objective accuracyImproved
Minkowski projection method vs path algorithm, ADMM, Gurobi on constrained lassozero-sum and nonnegative constrained lasso · synthetic, up to (n,d)=(8000,16000)runtimeImproved
Minkowski projection method vs ADMM sensitivity to sparsity levelconstrained lasso · lambda/lambda_max 0.2 to 0.8runtime sensitivityImproved
Viscosity penalty (rho>0) in block descentprojection onto polyhedral Minkowski sums (overlapping group lasso with p=infinity) · theoryconvergence rateImproved
Inexact projections (e.g., bisection for ball projection) in the block descentl1,p minimization · empirical remarkaccuracy/convergenceNo clear effect

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