AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Hierarchical Average Precision Training for Pertinent Image Retrieval

Elias Ramzi, Nicolas Audebert, Nicolas Thome, Clément Rambour, Xavier Bitot

ECCV 2022 · Vision-language and multimodal › Human and embodied modeling · approach: Composed and text-based image retrieval · 9 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
add smooth upper-bound surrogate loss instead of sigmoid (ablation)HAPPIER · iNat-baseH-AP: 52.3 -> 53.1Improved
add clustering loss (ablation)HAPPIER · iNat-baseH-AP: 53.1 -> 54.3Improved
train with H-AP instead of fine-level lossesResNet-50 · SOPH-AP: +16.1pt on SOPImproved
train with HAPPIER instead of CSLResNet-50 · SOPH-AP: +6pt on H-AP, +8pt on ASI and +2.6pts on NDCGImproved
train with HAPPIER instead of best hierarchical baselineResNet-50 · SOPH-AP: +6.3pt on SOP, +4.2pt on iNat-baseImproved
use HAPPIER instead of SigmaNSM on DyML-VehicleResNet-34 · DyML-VehiclemAP: 32.0 -> 37.0Improved
use HAPPIER instead of fine-level TL_SH on SOP coarse levelResNet-50 · SOPcoarse AP: 14.5 -> 58.4Improved
use HAPPIER instead of best fine-grained method at fine levelResNet-50 · SOPfine-level AP: 62.5 -> 60.4Hurt
raise alpha (HAPPIER_F) to favour fine level (ablation)HAPPIER · SOPfine-level AP: 60.4 -> 62.2Improved
raise alpha (HAPPIER_F) to favour fine level (ablation)HAPPIER · SOPcoarse-level AP: 58.4 -> 36.0Hurt
train with H-AP relevance instead of sum wAP (ablation)HAPPIER · iNat-baseH-AP: 52.0 -> 53.1Improved
increase lambda above 0.5 for clustering loss weight (ablation)HAPPIER · iNat-baseH-APHurt

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