AI Paper Atlas

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

MuCaLe-Net: Multi Categorical-Level Networks to Generate More Discriminating Features

Youssef Tamaazousti, Herve Le Borgne, Celine Hudelot

CVPR 2017 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 21 citations (OpenAlex)

Check 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
combine features from subordinate- and basic-level networks (ablation)MuCaLe-Net · VOC07 and NWOmAP: 70.3 -> 72.5; 51.2 -> 54.1Improved
add a superordinate-label network to subordinate and basic-level networks (ablation)MuCaLe-Net · VOC07 and NWOmAP: 72.5 -> 73.0; 54.1 -> 54.9Improved
use only basic-level labels instead of subordinate labels (ablation)MuCaLe-Net · VOC07 and NWOmAP: 70.3 -> 70.0; 51.2 -> 51.0Hurt
use only superordinate labels instead of subordinate labels (ablation)MuCaLe-Net · VOC07 and NWOmAP: 70.3 -> 58.9; 51.2 -> 37.4Hurt
train an additional network from a different random initializationstandard CNN descriptor · VOC07mAP: 70.3 -> 70.5±0.3No clear effect
randomly group specific categories for generic labelsstandard CNN descriptor · NWOmAP: 51.2 -> 51.0±0.3Hurt
jointly train on specific and generic labelsstandard CNN descriptor · VOC07 and NWOmAP: 70.3 -> 71.5; 51.2 -> 53.4Improved
use independently trained categorical-level networks and concatenate their featuresstandard CNN descriptor · VOC07 and NWOmAP: 70.3 -> 72.5; 51.2 -> 54.1Improved
use multi-categorical-level featuresVGG-16 transfer representation · Caltech-101, Caltech-256, VOC 2007, VOC 2012accuracy or mAP: 88.8 -> 92.0; 78.0 -> 80.9; 86.1 -> 87.5; 84.5 -> 86.1Improved

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