AI Paper Atlas

Research map / Video / Action recognition

Multi-Modal Few-Shot Temporal Action Segmentation

Zijia Lu, Ehsan Elhamifar

ICCV 2025 · Video › Action recognition · approach: Temporal action segmentation · 7 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 typed relational edges to the Action Relation Graph (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 8.9 -> 15.0Improved
use the dense graph instead of a sparse hierarchical graph (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 10.3 -> 15.0Improved
add task nodes to the graph (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 8.6 -> 15.0Improved
learn dynamic edge weights alpha instead of fixed alpha (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 12.1 -> 15.0Improved
use per-head alpha instead of shared alpha across heads (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 11.9 -> 15.0Improved
add prediction fusion between modalities (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 12.7 -> 15.0Improved
fuse predictions in the first SHCA layer instead of the last (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 13.5 -> 15.0Improved
add L2 penalty on influence factor tau_1 (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 11.1 -> 15.0Improved
use all three refinement stages instead of the first stage (ablation)PGNet (COIN, 3way-3shot) · COIN novel tasksF1@50: 11.4 -> 15.0Improved
add unlabeled support videos to zero-shot textual PGNetPGNet · COIN novel tasksF1@50: 8.6 -> 10.2Improved
add weak labels (one frame per action instance) to unlabeled few-shotPGNet · COIN novel tasksF1@50: 10.2 -> 13.8Improved
train on COIN instead of CrossTask base tasksPGNet (5way-3shot) · CrossTask novel tasksEdit and F1@50: Edit 34.7 -> 41.5; F1@50 17.5 -> 12.4Mixed

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