AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Contrastive Learning with Synthetic Positives

Dewen Zeng, Xinrong Hu, Yawen Wu, Xiaowei Xu, Yiyu Shi

ECCV 2024 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Inductive bias and minimal representations · 5 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 synthetic hard positives from diffusion feature interpolation to SimCLRSimCLR · CIFAR10 (linear)top-1 accuracy: 91.47 -> 94.37Improved
add synthetic hard positives to SimCLRSimCLR · CIFAR100 (linear)top-1 accuracy: 65.78 -> 72.01Improved
add synthetic positive to SimCLR on ImageNet100SimCLR ResNet-50 · ImageNet100top-1 accuracy: 75.75 -> 79.62Improved
replace augmented view with synthetic positive (no naive augmentation pair) (ablation)CLSP-SimCLR · CIFAR100 (linear)top-1 accuracy: 72.01 -> 54.46Hurt
replace augmented view with synthetic positive (no naive augmentation pair) (ablation)CLSP-SimCLR · CIFAR10 (linear)top-1 accuracy: 91.47 -> 92.86Improved
use synthetic positive instead of an extra naively augmented positive (ablation)CLSP-SimCLR · CIFAR100 (linear)top-1 accuracy: 68.30 -> 72.01Improved
increase positive candidate set size from 1 to 8 (ablation)CLSP-SimCLR · CIFAR100 (linear)top-1 accuracy: 65.84 -> 72.01Improved
increase candidate set size beyond 8 to 32 (ablation)CLSP-SimCLR · CIFAR100 (linear)top-1 accuracy: 72.01 -> 72.04No clear effect
increase loss weight lambda beyond 1.0 (ablation)CLSP-SimCLR · CIFAR100 (linear)top-1 accuracy: 72.01 -> 69.07Hurt
use larger feature interpolation weight w (ablation)CLSP-SimCLR · CIFAR10/CIFAR100linear accuracyHurt
use more than one additional synthetic positive per anchor (ablation)CLSP-SimCLR · CIFAR10top-1 accuracy: 94.37 -> 94.35No clear effect
use feature interpolation instead of RCG embedding-conditioned generation (ablation)CLSP-SimCLR · CIFAR100top-1 accuracy: 67.90 -> 72.01Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 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