AI Paper Atlas

Research map / Video / Video-language understanding

Rethinking Weakly-supervised Video Temporal Grounding From a Game Perspective

Xiang Fang, Zeyu Xiong, Wanlong Fang, Xiaoye Qu, Chen Chen, Jianfeng Dong, Keke Tang, Pan Zhou, Yu Cheng, Daizong Liu

ECCV 2024 · Video › Video-language understanding · approach: Temporal video grounding · 12 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 video self-modal game (ablation)word-level cross-modal game baseline · Charades-STAR@1, IoU=0.7: 19.34 -> 21.25Improved
add query self-modal game (ablation)video game + word-level · Charades-STAR@1, IoU=0.7: 21.25 -> 22.97Improved
add phrase-level cross-modal game (ablation)BA-Game with self games · Charades-STAR@1, IoU=0.7: 22.97 -> 24.81Improved
add sentence-level cross-modal game (ablation)full model · Charades-STAR@1, IoU=0.7: 24.81 -> 26.46Improved
add KLD soft supervision to the game header (ablation)game headers · Charades-STAR@1, IoU=0.7: 23.21 -> 26.46Improved
use KLD instead of MSE or CE for soft supervision (ablation)game headers · Charades-STAR@1, IoU=0.7: 24.73 (CE) -> 26.46Improved
use convolution layer plus self-attention header (ablation)game headers · Charades-STAR@1, IoU=0.7: 25.28 (LL) -> 26.46Improved
add reconstruction loss (ablation)full model · Charades-STAR@1, IoU=0.7: 24.35 -> 26.46Improved
replace Banzhaf index with Shapley index (ablation)game framework · Charades-STAR@1, IoU=0.7: 26.46 -> 26.28No clear effect
replace proposal-based matching with frame-wise game scoresWS-VTG · Charades-STAR@5, IoU=0.5: 84.71 -> 88.93 vs CPLImproved
replace proposal-based matching with game-based frame scoring in cross-dataset transferWS-VTG · Cross-dataset C->AR@1, IoU=0.5: 11.59 -> 20.37 vs CPLImproved
game-based training with coalition sampling versus CPLtraining cost · training speed per sampleseconds: 0.27s -> 0.94sHurt

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