AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Boosting First-Order Methods by Shifting Objective: New Schemes with Faster Worst-Case Rates

Kaiwen Zhou, Anthony Man-Cho So, James Cheng

NeurIPS 2020 · Optimization › Nonconvex and smooth optimization · approach: Adaptive first-order convex methods · 4 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
minimize shifted objective in BS-Point-SAGA instead of Point-SAGAPoint-SAGA · worst-case rate factor, large kappaexpected iteration complexity: almost 2 times lowerImproved
use G-TM (shifted objective) instead of Triple MomentumTriple Momentum (TM) · quadratic simulationconvergence rateNo clear effect
use G-TM instead of TM in the first iterationsTriple Momentum (TM) · l2-logistic regression on LIBSVM datasetsfunction value suboptimality: first 30 iterationsImproved
use G-TM or TM instead of NAGNesterov accelerated gradient · l2-logistic regression on LIBSVM datasetsfunction value suboptimalityImproved
use BS-SVRG instead of SAGA and Katyushastochastic first-order methods · a9a, l2-logistic regressionfunction value suboptimality vs data passesMixed
output x-tilde instead of z for BS-SVRG (ablation)BS-SVRG · a9aearly-stage performanceImproved
use numerical parameter choice (BS-SVRG-N) instead of analytic choice (ablation)BS-SVRG · l2-logistic regressionfunction value suboptimalityNo 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