AI Paper Atlas

Research map / Bandits and online learning / Online convex optimization and games

Conic Blackwell Algorithm: Parameter-Free Convex-Concave Saddle-Point Solving

Julien Grand-Clément, Christian Kroer

NeurIPS 2021 · Bandits and online learning › Online convex optimization and games · approach: Last-iterate convergence in games · 1 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
use CBA+ instead of OMD/FTRL variants at theoretical step size (alpha=1)CBA+ · DRO logistic classification (synthetic, splice, madelon)worst-case loss vs stepsImproved
multiply baseline step size by alpha=1,000OMD/FTRL baselines · DRO synthetic dataworst-case loss vs stepsMixed
multiply baseline step size by alpha=10,000OMD/FTRL baselines · DRO synthetic and spliceworst-case loss vs stepsHurt
use CBA+ instead of RM+ on matrix games and EFGsCBA+ · matrix games, Leduc, Battleshipduality gap vs stepsNo clear effect
use CBA+ instead of AdaHedge and AdaFTRLCBA+ · synthetic matrix gamesduality gap vs stepsImproved
use CBA+ instead of OMD, FTRL variants on DRO wall-clockCBA+ · DRO instancerunning time: 2x-2.5x faster; 3x-4x fasterImproved

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