AI Paper Atlas

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

Integrating Prior Knowledge in Contrastive Learning with Kernel

Benoit Dufumier, Carlo Alberto Barbano, Robin Louiset, Edouard Duchesnay, Pietro Gori

ICML 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 1 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
Replace InfoNCE with Decoupled UniformityResNet18 contrastive learning · CIFAR-10linear evaluation accuracy (%): 82.18_{\pm 0.30} -> \textbf{85.05}_{\pm 0.37}Improved
Replace InfoNCE with Decoupled UniformityResNet18 contrastive learning · CIFAR-100linear evaluation accuracy (%): 55.11_{\pm 0.20} -> \textbf{58.41}_{\pm 0.05}Improved
Replace InfoNCE with Decoupled UniformityResNet50 contrastive learning · ImageNet100linear evaluation accuracy (%): 68.76 -> 77.18Improved
Add BigBiGAN features as a kernel priorDecoupled Uniformity · ImageNet100linear evaluation accuracy (%): 77.18 -> 78.02Improved
Use weak attributes as kernel prior rather than conditional positive samplingDecoupled Uniformity versus CCLK · CUBlinear evaluation accuracy (%): 30.49 -> 32.87Improved
Use weak attributes as kernel prior rather than conditional positive samplingDecoupled Uniformity versus CCLK · ImageNet100linear evaluation accuracy (%): 77.24 -> 79.02Improved
Use weak attributes as kernel prior rather than conditional positive samplingDecoupled Uniformity versus CCLK · UT-Zapposlinear evaluation accuracy (%): 85.48 -> 85.56Improved
Add GloRIA features as a kernel priorDecoupled Uniformity · CheXpert, CardiomegalyAUC (%): 86.39 -> 85.88Hurt
Add GloRIA features as a kernel priorDecoupled Uniformity · CheXpert, ConsolidationAUC (%): 92.45 -> 93.03Improved
Add a VAE representation as kernel priorDecoupled Uniformity · BIOBD, BD vs HClinear evaluation AUC (%): 60.46_{\pm 1.23} -> \mathbf{62.19}_{\pm 1.58}Improved
Use a VAE kernel prior when adding shared noisy bits to viewsRandBits-CIFAR10 representation learning · RandBits-CIFAR10, 10 bitslinear evaluation accuracy (%): 13.67 -> \textbf{68.42}_{\pm 0.51}Improved
Use a VAE kernel prior when adding shared noisy bits to viewsRandBits-CIFAR10 representation learning · RandBits-CIFAR10, 20 bitslinear evaluation accuracy (%): 10.07 -> \textbf{68.58}_{\pm 0.17}Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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