AI Paper Atlas

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

Metric Learning With HORDE: High-Order Regularizer for Deep Embeddings

Pierre Jacob, David Picard, Aymeric Histace, Edouard Klein

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 73 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 HORDE regularizercontrastive loss (GoogleNet) · CUB-200-2011Recall@1 (%): 55.0 -> 57.1Improved
add HORDE regularizercontrastive loss (GoogleNet) · Cars-196Recall@1 (%): 72.2 -> 76.2Improved
add HORDE regularizerTriplet loss (GoogleNet) · CUB-200-2011Recall@1 (%): 50.5 -> 53.6Improved
add HORDE regularizerTriplet loss (GoogleNet) · Cars-196Recall@1 (%): 65.2 -> 74.0Improved
add HORDE regularizerBinomial Deviance (GoogleNet) · CUB-200-2011Recall@1 (%): 55.9 -> 58.3Improved
add HORDE regularizerBinomial Deviance (GoogleNet) · Cars-196Recall@1 (%): 78.2 -> 81.5Improved
add HORDE regularizerBinomial Deviance (GoogleNet) · Stanford Online ProductsRecall@1 (%): 67.4 -> 72.6Improved
add HORDE regularizerBinomial Deviance (GoogleNet) · In-Shop Clothes RetrievalRecall@1 (%): 81.3 -> 84.4Improved
add HORDE regularizerABE ensemble (GoogleNet) · CUB-200-2011Recall@1 (%): 60.0 -> 62.7Improved
add HORDE regularizerABE ensemble (GoogleNet) · Cars-196Recall@1 (%): 79.2 -> 86.4Improved
use all concatenated high-order moments at test time (PCA to 512) (ablation)Binomial Deviance + HORDE · CUB-200-2011Recall@1 (%): 58.3 -> 59.4Improved
use all concatenated high-order moments at test time (PCA to 512) (ablation)Binomial Deviance + HORDE · Cars-196Recall@1 (%): 81.5 -> 83.2Improved

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