AI Paper Atlas

Research map / Video / Video-language understanding

EgoDistill: Egocentric Head Motion Distillation for Efficient Video Understanding

Shuhan Tan, Tushar Nagarajan, Kristen Grauman

NeurIPS 2023 · Video › Video-language understanding · approach: Egocentric video understanding · 6 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
remove L_KD (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 39.33Hurt
remove L_1 feature-matching loss (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 42.20Hurt
remove L_GT (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 43.17Hurt
remove L_KD (ablation)EgoDistill · Ego4Dtop-1 accuracy: 37.95 -> 35.51Hurt
remove self-supervised IMU pretraining (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 41.21Hurt
remove L_1 pretraining (keep L_NCE) (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 42.30Hurt
remove L_NCE pretraining (keep L_1) (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 43.51Hurt
add IMU and distillation to image model trained with L_GT only (ablation)VisOnly · EPIC-Kitchenstop-1 accuracy: 35.04 -> 44.95Improved
add IMU branch to VisOnly-DistillVisOnly-Distill (MotionFormer teacher) · EPIC-Kitchenstop-1 accuracy: 37.20 -> 44.95Improved
add IMU branch to VisOnly-DistillVisOnly-Distill (SlowFast teacher) · Ego4Dtop-1 accuracy: 33.04 -> 33.29Improved
use MViT teacher instead of MotionFormer teacherEgoDistill · EPIC-Kitchenstop-1 accuracy: 44.95 -> 36.90Hurt
uniform vs last-frame sampling (N=4) (ablation)EgoDistill · EPIC-Kitchenstop-1 accuracy: 41.72 -> 52.43Improved

Later papers that compared with this method

1 later accepted paper reports EgoDistill as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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