AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Generalized Ambiguity Decomposition for Ranking Ensemble Learning

Hongzhi Liu, Yingpeng Du, Zhonghai Wu

JMLR 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Recommendation representation learning

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
add list-wise loss in place of pair-wise loss (ablation)weighted ranking ensemble · Amazon movieRecall@5: 0.0418 -> 0.0555Improved
add list-wise loss in place of pair-wise loss (ablation)weighted ranking ensemble · MQ2007Recall@5: 0.3081 -> 0.3097No clear effect
add diversity-based learning to WEL (ablation)WEL · MQ2007Recall@5: 0.3097 -> 0.3162Improved
add diversity-based learning to WEL (ablation)WEL · MQ2008Recall@5: 0.7022 -> 0.7007Hurt
add diversity term to aWEL (ablation)aWEL · ML-20mRecall@5: 0.3675 -> 0.3695Improved
add adaptive query-specific weights to WELv (ablation)WELv · MQ2007Recall@5: 0.3162 -> 0.3497Improved
add adaptive query-specific weights to WEL (ablation)WEL · MQ2008Recall@5: 0.7022 -> 0.7471Improved
use aWELv in place of best baseline AEMranking ensembles · EpinionRecall@5: 0.0710 -> 0.0794Improved
use too large or too small diversity trade-off lambda (ablation)aWELv · RS and IR datasetsRecallHurt
vary interpolation constant c from 0 to 1 (ablation)aWELv · RS and IR datasetsRecallNo clear effect
ensemble simple base models instead of using a single state-of-the-art modelaWELv · RS tasks vs deep modelsRecall@20: 0.414 -> 0.424Improved
add diversity term Div to list-wise ensemble (ablation)WEL · training vs test curvesoverfittingImproved

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