AI Paper Atlas

Research map / Video / Video-language understanding

Contrastive Video Question Answering via Video Graph Transformer

Junbin Xiao, Pan Zhou, Angela Yao, Yicong Li, Richang Hong, Shuicheng Yan, Tat‐Seng Chua

TPAMI 2023 · Video › Video-language understanding · approach: Video question answering · 75 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 dynamic graph transformer (DGT) instead of mean pooling (ablation)VGT · TGIF-QA Actionaccuracy: 89.6 -> 95.0Improved
add node transformer (ablation)VGT · NExT-QA valAcc@All: 53.84 -> 55.02Improved
add edge transformer (ablation)VGT · NExT-QA valAcc@All: 54.30 -> 55.02Improved
add temporal graph transformer (ablation)VGT · STAR ValAcc@M: 42.37 -> 44.27Improved
add frame-level feature F_I to object graphs (ablation)VGT · NExT-QA valAcc@All: 53.58 -> 55.02Improved
contrastive answer matching instead of [CLS] classification (ablation)VGT · NExT-QAAcc@All: 45.82 -> 55.02Improved
add self-supervised contrastive loss L_vq (ablation)VGT (BERT) · NExT-QAAcc@All: 55.02 -> 57.15Improved
add self-supervised contrastive loss L_vq (ablation)VGT (BERT) · TGIF-FrameQAaccuracy: 61.6 -> 61.3No clear effect
replace BERT with RoBERTa (no L_vq) (ablation)VGT · NExT-QA / STAR-QAAcc@All / Acc@M: 55.02 -> 58.45 (NExT-QA); 44.27 -> 42.44 (STAR-QA)Mixed
add 4 negative questions in L_vq (ablation)CoVGT · NExT-QAaccuracy: 58.45 -> 60.01Improved
hard negatives by question type instead of random (ablation)CoVGT · NExT-QAaccuracy: 59.35 -> 60.01Improved
pretrain on 0.18M WebVid data (ablation)VGT · NExT-QA valAcc@All: 55.02 -> 56.02Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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