AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

Focal Visual-Text Attention for Memex Question Answering

Junwei Liang, Lu Jiang, Liangliang Cao, Yannis Kalantidis, Li-Jia Li, Alexander G. Hauptmann

TPAMI 2019 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA · 71 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 FVTA kernel tensor instead of cosine similarity (ablation)FVTA · MemexQAaccuracy: 0.619 -> 0.669Improved
add intra-sequence dependency (ablation)FVTA · MemexQAaccuracy: 0.569 -> 0.669Improved
add cross-sequence interaction (ablation)FVTA · MemexQAaccuracy: 0.604 -> 0.669Improved
add question attention (ablation)FVTA · MemexQAaccuracy: 0.629 -> 0.669Improved
add photo (visual) input (ablation)FVTA · MemexQAaccuracy: 0.577 -> 0.669Improved
use FVTA attention instead of soft attentionFVTA · MemexQAaccuracy: 0.621 -> 0.669Improved
use FVTA instead of TGIF temporal attentionFVTA · MemexQAaccuracy: 0.630 -> 0.669Improved
use attention instead of no attention embedding methodsFVTA · MemexQAaccuracy: 0.563 -> 0.669Improved
use FVTA instead of TGIF temporal on when questionsFVTA · MemexQA (when questions)accuracy: 0.522 -> 0.476Hurt
use FVTA attention for evidence localization instead of TGIFFVTA · MemexQA evidence groundingHIT@1: 13.28% -> 15.48%Improved
use FVTA instead of RWMNFVTA · MovieQA valaccuracy: 0.387 -> 0.410Improved
use FVTA instead of RWMNFVTA · MovieQA testaccuracy: 0.363 -> 0.373Improved

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