AI Paper Atlas

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

Question Type Guided Attention in Visual Question Answering

Yang Shi, Tommaso Furlanello, Sheng Zha, Animashree Anandkumar

ECCV 2018 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA · 55 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 QTA weighting of image features (ablation)CATL · TDIUC testoverall accuracy: 82.23 -> 83.62Improved
add QTA weighting of image features (ablation)CATL (W) · TDIUC testoverall accuracy: 83.92 -> 85.03Improved
add QTA weighting of image features (ablation)CATL (W) · TDIUC testharmonic MPT: 57.83 -> 60.08Improved
add QTA weighting of image features (ablation)CATL · TDIUC testharmonic MPT: 55.95 -> 54.60Hurt
concatenate ResNet and Faster R-CNN features (ablation)CAT1 · TDIUC testoverall accuracy: 72.44 -> 73.05Improved
use Faster R-CNN instead of ResNet features (ablation)CAT1 with NMT · TDIUC testoverall accuracy: 71.41 -> 72.44Improved
replace pre-trained NMT text feature with jointly-trained LSTM (ablation)CAT1 with ResNet · TDIUC testoverall accuracy: 71.41 -> 77.55Improved
add Word2Vec to LSTM text feature (ablation)CAT1L Faster R-CNN · TDIUC testoverall accuracy: 78.66 -> 80.16Improved
add question type as plain embedding input (QT) instead of attention (QTA) (ablation)CATL / CAT · TDIUC testaccuracy: 3-5%Improved
add absurd-question removal in training (ablation)CATL-QTA-M · TDIUC testsimple accuracy: 84.33 -> 80.95Mixed
add QTA to MCB modelMCB · TDIUC testoverall accuracy: 84.97 vs 85.03Hurt
predict question type with multi-task head instead of ground-truth type (ablation)CATL-QTA (W) · TDIUC testsimple accuracy: 85.03 -> 84.33Hurt

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