AI Paper Atlas

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

Instance Similarity Learning for Unsupervised Feature Representation

Ziwei Wang, Yunsong Wang, Ziyi Wu, Jiwen Lu, Jie Zhou

ICCV 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 9 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 hard positive enhancement (ablation)ISL AlexNet · CIFAR-10 (weighted kNN, FC)top-1 accuracy (%): 81.1 -> 82.8Improved
add hard positive enhancement (ablation)ISL ResNet18 · CIFAR-100 (weighted kNN, FC)top-1 accuracy (%): 52.1 -> 54.7Improved
add hard positive enhancement (ablation)ISL ResNet50 · CIFAR-100 (weighted kNN, FC)top-1 accuracy (%): 56.7 -> 58.1Improved
add hard positive enhancement (ablation)ISL ResNet50 · CIFAR-100 (linear, conv5)top-1 accuracy (%): 63.0 -> 65.9Improved
add hard positive enhancement (ablation)ISL AlexNet · SVHN (weighted kNN, FC)top-1 accuracy (%): 91.0 -> 91.8Improved
use instance similarity learning without HPE instead of AND neighborhood discoveryISL w/o HPE AlexNet · CIFAR-100 (weighted kNN, FC)top-1 accuracy (%): 41.5 -> 49.2Improved
use ISL instead of ANDResNet50 · CIFAR-100 (weighted kNN, FC)top-1 accuracy (%): 49.0 -> 58.1Improved
use ISL instead of RotNetAlexNet · CIFAR-10 (linear, conv5)top-1 accuracy (%): 84.1 -> 85.8Improved
use ISL instead of LAResNet50 · ImageNet (weighted kNN, FC)top-1 accuracy (%): 49.4 -> 50.2Improved
integrate ISL with MoCo-v2MoCo-v2 ResNet50 · ImageNet (linear, conv5)top-1 accuracy (%): 67.5 -> 68.6Improved
integrate ISL with MoCo-v2MoCo-v2 ResNet50 · ImageNet (weighted kNN, FC)top-1 accuracy (%): 38.5 -> 40.1Improved
integrate ISL with MoCo-v2MoCo-v2 ResNet50 · ImageNet (linear, conv4)top-1 accuracy (%): 52.9 -> 51.7Hurt

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