AI Paper Atlas

Research map / Video / Action recognition

Composable Augmentation Encoding for Video Representation Learning

Chen Sun, Arsha Nagrani, Yonglong Tian, Cordelia Schmid

ICCV 2021 · Video › Action recognition · approach: Contrastive video representation learning · 18 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 time shift encoding to the transformer head (ablation)SimCLR++ · SSv1 linear evaltop-1 accuracy: 26.5 -> 31.2Improved
add temporal augmentation (frames at different times) to MLP-head SimCLR (ablation)SimCLR · SSv1 linear evaltop-1 accuracy: 17.1 -> 26.4Improved
add MLP head on top of the transformer head without encodings (ablation)SimCLR++ · SSv1 linear evaltop-1 accuracy: 26.5 -> 24.6Hurt
add crop encoding with dropout (ablation)SimCLR++ · SSv1 linear evaltop-1 accuracy: 26.5 -> 28.1Improved
add dropout on crop encoding (ablation)CATE crop encoding · SSv1 linear evaltop-1 accuracy: 27.2 -> 28.1Improved
encode time shift delta_t instead of only arrow of time (ablation)CATE time encoding · SSv1 linear evaltop-1 accuracy: 28.1 -> 31.3Improved
add dropout on time encoding (ablation)CATE time encoding · SSv1 linear evaltop-1 accuracy: 31.3 -> 31.2No clear effect
compose crop encoding with time encoding (ablation)CATE · SSv1 linear evaltop-1 accuracy: 31.2 -> 32.2Improved
compose crop encoding with time encoding (ablation)CATE · SSv2 linear evaltop-1 accuracy: 42.3 -> 43.5Improved
use 4 transformer layers instead of 1 (ablation)CATE · SSv1 linear evaltop-1 accuracy: 26.9 -> 31.2Improved
use 8 transformer layers instead of 4 (ablation)CATE · SSv1 linear evaltop-1 accuracy: 31.2 -> 31.3No clear effect
pretrain 800 epochs instead of 500 (ablation)CATE · SSv1, UCF101, HMDB51accuracy: 32.2 -> 33.1; 84.3 -> 83.6; 53.6 -> 53.0Mixed

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