AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Bayesian Triplet Loss: Uncertainty Quantification in Image Retrieval

Frederik Warburg, Martin Jørgensen, Javier Civera, Søren Hauberg

ICCV 2021 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Priors and functional variational uncertainty · 48 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
use Gaussian Bayesian triplet loss instead of triplet regressionTriplet loss, ResNet50 · CUB200ECE@5: 0.331 -> 0.037Improved
use Bayesian triplet loss instead of plain triplet lossResNet50 · CUB200Recall@1: 0.648 -> 0.612Hurt
use expected distance instead of Euclidean distance for nearest neighbor searchBayes Triplet, ResNet50 · CUB200Recall@1: 0.612 -> 0.632Improved
use expected distance for search with Densenet161Bayes Triplet, Densenet161 · CUB200Recall@1: 0.683 -> 0.617Hurt
use larger Densenet161 backbone instead of ResNet50Triplet · CUB200Recall@1: 0.648 -> 0.717Improved
add MC dropout (MC=50) to triplet lossTriplet, Densenet161 · CUB200Recall@1: 0.717 -> 0.349Hurt
add MC dropout (MC=50) to triplet lossTriplet, Densenet161 · MSLSRecall@1: 0.386 -> 0.282Hurt
use von Mises-Fisher embeddings instead of triplet lossResNet50 · Car-196 (OOD)Recall@1: 0.451 -> 0.471Improved
use Gaussian Bayesian triplet loss instead of triplet regressionDensenet161 · Car-196 (OOD)ECE@5: 0.456 -> 0.134Improved
use Gaussian Bayesian triplet loss instead of triplet regressionDensenet161 · MSLSECE@5: 0.543 -> 0.264Improved
use Gaussian Bayesian triplet loss instead of triplet lossResNet50 · MSLSRecall@1: 0.350 -> 0.354No clear effect

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