AI Paper Atlas

Research map / Video / Action recognition

Minority-Oriented Vicinity Expansion with Attentive Aggregation for Video Long-Tailed Recognition

WonJun Moon, Hyun Seok Seong, Jae-Pil Heo

AAAI 2023 · Video › Action recognition · approach: Contrastive video representation learning · 3 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 the two learnable feature aggregators (ablation)VideoLT baseline · VideoLTAll accuracy: 0.516 -> 0.690Improved
Add the self-attentive aggregator (ablation)VideoLT baseline · VideoLTAll accuracy: 0.516 -> 0.681Improved
Add the codebook-attentive aggregator (ablation)VideoLT baseline · VideoLTAll accuracy: 0.516 -> 0.677Improved
Add both aggregators rather than either one alone (ablation)VideoLT baseline · VideoLTAll accuracy: 0.681 / 0.677 -> 0.690Improved
Add dynamic extrapolation (ablation)Both aggregators on VideoLT · VideoLTTail accuracy: 0.601 -> 0.632Improved
Add calibrated interpolation (ablation)Both aggregators on VideoLT · VideoLTTail accuracy: 0.601 -> 0.632Improved
Combine dynamic extrapolation and calibrated interpolation (ablation)Both aggregators on VideoLT · VideoLTAll accuracy: 0.690 -> 0.719Improved
Use prototypical self-attention instead of conventional MSA (ablation)VideoLT with TSM features · VideoLTAll accuracy: 0.642 -> 0.669Improved
Use actual class sample counts rather than a linear tail-weight criterion (ablation)MOVE · VideoLTAll accuracy: 0.699 -> 0.719Improved
Use MOVE instead of the strongest reported baselineVideoLT with ResNet-101 features · VideoLTTail accuracy: 0.611 -> 0.644Improved
Use MOVE instead of MixupImbalanced-MiniKinetics200 with ResNet-50, imbalance ratio 0.01 · Imbalanced-MiniKinetics200accuracy: 0.488 -> 0.509Improved
Use Mixup rather than MOVEImbalanced-MiniKinetics200 with ResNet-50, imbalance ratio 0.01 · Imbalanced-MiniKinetics200AP: 0.570 -> 0.570No clear effect

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