AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Cost efficient gradient boosting

Sven Peter, Ferran Diego, Fred A. Hamprecht, Boaz Nadler

NeurIPS 2017 · Learning theory and algorithms › Prediction and decision losses · approach: Tree ensemble optimization · 36 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
Best-first (leaf-wise) tree growth instead of breadth-first growth, with leaf count limit (ablation)CEGB tree construction · Yahoo! LTR (fixed cost budget)Average Precision@5 at given prediction costImproved
Shallow trees (16 leaves, depth 4) under CEGB instead of deep trees (ablation)CEGB · Yahoo! LTR (feature + evaluation cost)cost-accuracy tradeoff vs GreedyMiserImproved
Allowing deep branches (unrestricted depth, leaf-limited) in CEGBCEGB vs GreedyMiser · Yahoo! LTR (feature + evaluation cost)accuracy at prediction costImproved
Training with uniform feature cost (beta=1) and applying original costs only at test time (ablation)CEGB · Yahoo! LTRcost-accuracy tradeoffNo clear effect
Large feature-cost penalty lambda (0.001 is too small/ lower accuracy); larger lambda reaches similar accuracy at much cheaper cost (ablation)CEGB · Yahoo! LTRaccuracy vs costMixed
Multi-scale features (coarse-resolution features shared across pixel blocks) with evaluation-cost-dominated settingCEGB · Flies video multi-scale classification (SEGMeasure)evaluation cost at similar accuracyImproved
CEGB vs GreedyMiser and BudgetPrune under equal-contribution costs on MiniBooNE, Covertype, HEPMASS, Yahoo binarizedCEGB · Multiple datasetscost-accuracy tradeoffImproved

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