AI Paper Atlas

Research map / Optimization / Optimizers

Hierarchical Optimization-Derived Learning

Risheng Liu, Xuan Liu, Shangzhi Zeng, Jin Zhang, Yixuan Zhang

TPAMI 2023 · Optimization › Optimizers · approach: Constrained and non-backprop training

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 HODL hierarchical training instead of ENA (5 layers)ENA · Set14 constrained sparse codingPSNR: 15.59 -> 18.82Improved
use HODL instead of UNH (5 layers)UNH · Set14 constrained sparse codingPSNR: 10.47 -> 18.82Improved
increase layers from 5 to 25 (ablation)HODL · Set14 constrained sparse codingPSNR: 18.82 -> 18.98No clear effect
increase layers from 5 to 25 (ablation)ENA · Set14 constrained sparse codingPSNR: 15.59 -> 15.64No clear effect
increase layers from 5 to 25 (ablation)UNH · Set14 constrained sparse codingPSNR: 10.47 -> 11.31Improved
use ALM instead of ADMM as base solver (ablation)ENA · Set14 constrained sparse codingPSNR: 15.58 -> 13.78Hurt
use ALM instead of ADMM as base solver (ablation)UNH · Set14 constrained sparse codingPSNR: 11.27 -> 7.32Hurt
HODL with ALM versus ENA with ADMMHODL · Set14 constrained sparse codingPSNR: 15.58 -> 18.64Improved
add non-expansive property of the operator (ablation)HODL · constrained sparse coding convergencegradient of learning objective: reduces the gradient of the learning objective by an order of magnitudeImproved
use aggregation (mu greater than 0) instead of sHODL (mu equals 0) (ablation)aHODL · constrained sparse coding convergenceconvergence of optimization variableMixed
drop aggregation to use simplified sHODL (ablation)aHODL · constrained sparse codingcomputational costImproved
use HODL instead of RCDNet on Rain100LRCDNet · Rain100LPSNR: 40.00 -> 40.07No 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