AI Paper Atlas

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

Spatially Aware Multimodal Transformers for TextVQA

Yash Kant, Dhruv Batra, Peter Anderson, Alexander Schwing, Devi Parikh, Jiasen Lu, Harsh Agrawal

ECCV 2020 · Vision-language and multimodal › Visual question answering · approach: Multimodal LLM question answering · 97 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
replace Rosetta-en OCR with Google OCRM4C · TextVQA valVQA accuracy: 39.4 -> 41.8Improved
replace ResNet-101 with ResNeXt-151 detector backboneM4C · TextVQA valVQA accuracy: 41.8 -> 42.0Improved
add two more transformer layers (6N)M4C · TextVQA valVQA accuracy: 42.0 -> 42.7Improved
jointly train on ST-VQAM4C improved · TextVQA valVQA accuracy: 42.7 -> 43.3Improved
use beam search (size 5)M4C improved · TextVQA valVQA accuracy: 43.3 -> 43.8Improved
replace 4 of 6 self-attention layers with spatially aware attention (ablation)M4C improved · TextVQA valVQA accuracy: 42.70 -> 43.80Improved
replace only 1 of 6 layers with spatially aware attention (ablation)M4C improved · TextVQA valVQA accuracy: 42.70 -> 42.61No clear effect
replace 5 of 6 layers (leaving one normal layer) instead of 4 (ablation)SA-M4C · TextVQA valVQA accuracy: 43.80 -> 43.07Hurt
set context size c=2 instead of 1 (ablation)SA-M4C · TextVQA valVQA accuracy: 43.80 -> 43.90Improved
use spatial graph masking instead of random masking (ablation)SA-M4C · TextVQA valVQA accuracy: 42.09 -> 43.80Improved
use spatial graph instead of top-k attention (ablation)SA-M4C · TextVQA valVQA accuracy: 43.26 -> 43.90Improved
use spatially aware attention instead of ReGAT graph attention (ablation)SA-M4C · TextVQA valVQA accuracy: 43.20 -> 43.90Improved

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