AI Paper Atlas

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

Attention-based Ensemble for Deep Metric Learning

Wonsik Kim, Bhavya Goyal, Kunal Chawla, Jungmin Lee, Keunjoo Kwon

ECCV 2018 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 238 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 divergence loss to ABE-8 (ablation)ABE-8 (512) · CARS-196Recall@1 (%): 69.7 -> 85.2Improved
add divergence loss to 8-heads ensemble (ablation)8-heads · CARS-196Recall@1 (%): 76.1 -> 76.0No clear effect
add attention modules to 8-heads (ablation)8-heads · CARS-196Recall@1 (%): 76.1 -> 79.7Improved
share global embedding G across learners and add divergence loss on top of attention (ablation)8-heads+att · CARS-196Recall@1 (%): 79.7 -> 85.2Improved
use shared G with separate S_m (8-tails) instead of 8-heads (ablation)8-heads · CARS-196Recall@1 (%): +5.0%Improved
add attention modules to 8-tails (ablation)8-tails · CARS-196Recall@1 (%): 81.1 -> 85.2Improved
increase number of attention learners M from 1 to 8 (ablation)ABE-M (512) · CARS-196Recall@1 (%): 67.3 -> 85.2Improved
replace 4-heads ensemble with ABE-4 attention ensemble4-heads (512) · CARS-196Recall@1 (%): 76.6 -> 82.5Improved
increase M with fixed 64-d individual learner embedding (ablation)ABE-M individual learners · CARS-196Recall@1 (%): 65.9 -> 75.0Improved
increase M in M-heads from 4 to 8 heads (ablation)M-heads · CARS-196Recall@1 (%): 76.6 -> 76.1No clear effect
use 5 inception blocks in attention module (4a to 4e) (ablation)ABE-8 · CARS-196Recall@1 (%)Improved
branch attention at pool3 (ablation)ABE-8 · CARS-196Recall@1 (%)Improved

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