AI Paper Atlas

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

VQA: Visual Question Answering

Aishwarya Agrawal, Jiasen Lu, Stanislaw Antol, Margaret Mitchell, C. Lawrence Zitnick, Devi Parikh, Dhruv Batra

IJCV 2016 · Vision-language and multimodal › Visual question answering · approach: Universal VQA models · 4,730 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 l2 normalization of image features (ablation)deeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 57.59 -> 57.75Improved
fuse by element-wise multiplication instead of concatenation (ablation)deeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 56.80 -> 57.75Improved
use 1000 answers instead of 500 (ablation)deeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 56.93 -> 57.75Improved
use 2000 answers instead of 1000deeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 57.75 -> 58.15Improved
truncate question vocabulary at 5 occurrencesdeeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 57.75 -> 57.99Improved
truncate question vocabulary at 11 occurrencesdeeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 57.75 -> 57.81No clear effect
train on filtered dataset keeping confident answersdeeper LSTM Q + norm I · VQA real images test-dev (open-ended)accuracy: 57.75 -> 56.62Hurt
add image features to question features (BoW Q + I)BoW Q · VQA real images test-dev (open-ended)accuracy: 48.09 -> 52.64Improved
add image to LSTM QLSTM Q · VQA real images test-dev (open-ended)accuracy: 48.76 -> 53.74Improved
use a two-layer LSTM with normalized image featuresLSTM Q + I · VQA real images test-dev (open-ended)accuracy: 53.74 -> 57.75Improved
add human captions to BoW questionBoW Q · VQA real images val (open-ended)accuracy: 48.09 -> 54.70Improved
add image features for reasoning-heavy question types (how many)Question-only model · VQA question types (open-ended)accuracy: 43.78 -> 40.33Hurt

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