AI Paper Atlas

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

Learning Compact Model for Large-Scale Multi-Label Data

Tong Wei, Yu-Feng Li

AAAI 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Partial and multi-label learning · 10 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
apply POP label and feature parameter pruning to Binary RelevanceBR · wiki10model size: 23.50 GB -> 67.50 MBImproved
apply POP pruning to BRBR · eurlexmodel size: 156.38 MB -> 20.18 MBImproved
apply POP pruning to BRBR · eurlexPSP@1: 39.93 -> 40.06No clear effect
apply POP pruning to BRBR · bibtexPSP@5: 59.34 -> 58.86No clear effect
apply POP pruning to BRBR · wiki10PSP@5: 13.96 -> 13.46No clear effect
add label parameter pruning before feature thresholding (POP vs DiSMEC feature-only pruning) (ablation)POP · wiki10model size: 880.00 MB -> 67.50 MBImproved
add label parameter pruning (POP vs DiSMEC) (ablation)POP · eurlexmodel size: 79.86 MB -> 20.18 MBImproved
add label parameter pruning (POP vs DiSMEC) (ablation)POP · eurlexPSP@1: 41.20 -> 40.06No clear effect
increase tolerance epsilon (ablation)POP · delicious, eurlex, wiki10P@kHurt
increase tolerance epsilon (ablation)POP · delicious, eurlex, wiki10model size reductionImproved
increase lambda too far (ablation)POP · delicious, eurlex, wiki10P@kHurt

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