AI Paper Atlas

Research map / Detection and segmentation / CNN features

LoCo: Local Contrastive Representation Learning

Yuwen Xiong, Mengye Ren, Raquel Urtasun

NeurIPS 2020 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 19 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 overlapping shared stage between local units (LoCo) to gradient-isolated local learning (ablation)SimCLR local learning · ImageNet-1K linear eval, 100 eptop-1 accuracy: 65.1 -> 66.2Improved
LoCo instead of GIM on SimCLRResNet-50 · ImageNet-1K linear evaltop-1 accuracy: 64.7 -> 69.5Improved
add one conv block with downsampling to local decoders (ablation)local contrastive learning · ImageNet-1K linear eval, 100 eptop-1 accuracy: 60.9 -> 65.1Improved
add one conv block to end-to-end SimCLR encoder end (ablation)SimCLR · ImageNet-1K linear eval, 100 eptop-1 accuracy: 65.7 -> 65.6No clear effect
remove downsampling from the added decoder conv block (ablation)local decoder with 1 conv block · ImageNet-1K linear eval, 100 eptop-1 accuracy: 65.1 -> 63.6Hurt
deepen local decoder to 2 conv blocks, 1 stage or full network (ablation)local learning without sharing · ImageNet-1K linear eval, 100 eptop-1 accuracy: 65.1 -> 65.8Improved
add two more MLP decoder layers (4 total) (ablation)local learning without sharing · ImageNet-1K linear eval, 100 eptop-1 accuracy: 60.9 -> 62.3Improved
add two more MLP decoder layers to LoCo (ablation)LoCo · ImageNet-1K linear eval, 100 eptop-1 accuracy: 66.2 -> 67.5Improved
keep only upper-module gradients on shared stage (ablation)LoCo · ImageNet-1K linear eval, 100 eptop-1 accuracy: 66.2 -> 65.3Hurt
replace hard sharing with L2 soft sharing (ablation)LoCo · ImageNet-1K linear eval, 100 eptop-1 accuracy: 66.2 -> 66.0Hurt
share only 1 block per stage instead of 1 stage (ablation)LoCo · ImageNet-1K linear eval, 100 eptop-1 accuracy: 66.2 -> 64.8Hurt
use Progressive ResNet-50 instead of ResNet-50 for local learning with LoCoLoCo · ImageNet-1K linear eval, 100 epmemory saving ratio: 1.28x -> 2.76xImproved

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